Carbon monoxide coordination and reversible photodissociation in copper(I) pyridylalkylamine compounds

Inorg. Chem. 2008, 47, 241−256

Carbon Monoxide Coordination and Reversible Photodissociation in

Copper(I) Pyridylalkylamine Compounds

H. Christopher Fry, Heather R. Lucas, Amy A. Narducci Sarjeant, Kenneth D. Karlin,* and

Gerald J. Meyer*

The Johns Hopkins UniVersity, Department of Chemistry, 3400 North Charles Street,

Baltimore, Maryland 21218

Received September 25, 2007

Systematic studies of CO coordination and photodissociation have been carried out for a series of copper(I) carbonyl

compounds possessing tripodal tetradentate ligands, [Cu^I(L)(CO)]B(C₆F₅)₄(L Me₂N-TMPA (1^Me2N), MeO-TMPA

)

(1^MeO), H-TMPA (1^H), PMEA (2^pmea), PMAP (2^pmap), BQPA (3^bqpa). Detailed structural, electrochemical, and infrared

characterization has been accomplished. In addition, various experimental techniques were utilized to determine

equilibrium binding constants (K_CO), association (k_CO), and dissociation (k_-CO) rate constants, as well as the

thermodynamic (

ligand-electron-donating ability, greater

on the order 1^Me2NCO > 1^MeOCO > 1^H

rings like 1^Rto six-membered chelate rings like 2^R, the stability of the CO adduct decreases, 1^H

> 2^pmapCO. The CO-binding properties of 3^bqpadid not follow trends observed for the other compounds, presumably

∆

H

°

,

∆S

°

) and activation parameters (

back-bonding results in stronger Cu

CO. With systematic synthetic expansion of the five-membered chelate

CO > 2^pmea

CO

∆

H^q,

∆

S^q) that regulate these processes. With increased

CO bonds, leading to K_COvalues

π

−

because of its bulkier ligand framework. Through solid- and solution-state analyses, we concluded that the photolabile

carbonyl species in solution possess a tridentate coordination mode, forming strictly five-membered chelate rings

to the copper ion with one dangling arm of the tripodal ligand. Carbon monoxide reversibly photodissociated from

complexes 1^Me2N

solvent but not from 2^pmea

as well as to small molecule (O₂, CO, NO) reactions with hemes and copper proteins are discussed.

−

CO, 1^MeO

−

CO, 1^H

−

CO, and 3^bqpa

−CO in coordinating (CH₃CN) and weakly coordinating (THF)

−

CO and 2^pmap

−CO. Comparisons to O₂-binding data available for these copper complexes

Introduction

metalloproteins, CO has been employed as a redox-inactive

surrogate for dioxygen, as its readily observable infrared (IR)

absorption band(s) can provide insights into the nature of

the copper coordination environment, small molecule as-

sociation/dissociation dynamics, and active-site locale.

Early work involving CO binding to copper proteins such

as hemocyanin^7-9(Hc, arthropodal and molluscan blood O₂-

carrier) aided in the description of the dicopper catalytic

center as unsymmetrical because of the unequal binding of

CO (1 equiv per dicopper). Spectroscopic and chemical

similarities to Hc show that the active sites of tyrosinase^10-12

(Tyr, ubiquitous o-phenol monooxygenase) and catechol

Carbon monoxide (CO) and dioxygen (O₂) are known to

competitively interact with iron- and copper-ion-containing

enzymes. Both ligands have empty π-antibonding orbitals

that are available for interactions with filled-metal d orbitals.

Copper(I) and its redox partner copper(II) comprise the

chemistry involved in copper metalloprotein function, such

as electron transfer¹or the processing of dioxygen (O₂)^2-5

and/or nitrogen oxides.⁶To probe the reduced form of such

* To whom correspondence should be addressed. karlin@jhu.edu

(K.D.K.), meyer@jhu.edu (G.J.M.).

(1) Rorabacher, D. B. Chem. ReV. 2004, 104, 651-697.

(2) Quant Hatcher, L.; Karlin, K. D. J. Biol. Inorg. Chem. 2004, 9, 669-

683.

(6) Wasser, I. M.; de Vries, S.; Moe¨nne-Loccoz, P.; Schro¨der, I.; Karlin,

K. D. Chem. ReV. 2002, 102, 1201-1234.

(3) Mirica, L. M.; Ottenwaelder, X.; Stack, T. D. P. Chem. ReV. 2004,

(7) Bonaventura, C.; Sullivan, B.; Bonaventura, J.; Bourne, S. Biochemistry

1974, 13, 4784-4789.

104, 1013-1045.

(4) Lewis, E. A.; Tolman, W. B. Chem. ReV. 2004, 104, 1047-1076.

(5) Lee, Y.; Karlin, K. D. In Concepts and Models in Bioinorganic

Chemistry; Metzler-Nolte, N., Kraatz, H.-B., Eds.; Wiley-VCH: New

York, 2006; pp 363-395.

(8) Sorrell, T. N.; Jameson, D. L. J. Am. Chem. Soc. 1982, 104, 2053-

2054.

(9) Ellerton, H. D.; Ellerton, N. F.; Robinson, H. A. Prog. Biophys. Mol.

Biol. 1983, 41, 143-248.

10.1021/ic701903h CCC: $40.75

Published on Web 12/04/2007

Inorganic Chemistry, Vol. 47, No. 1, 2008 241

Fry et al.

oxidase^13-15are also asymmetric. More recently, stoichio-

metric CO binding studies and IR spectroscopic measure-

ments helped to characterize the structural and functional

inequivalence of the two copper(I) centers (Cu_At Cu_H, Cu_B

t Cu_M) of reduced dopamine-â-monoxygenase (DâH)^16-18

and peptidylglycine monooxygenase (PHM).^19-22In addition,

CO facilitated the characterization/description of an unprec-

edented S-donor methionine ligand at the Cu_B/Cu_Msite and

aided in the elucidation of the ligand environment for the

Cu^Iversus Cu^IIcenters.

Chart 1

Although CO has been most commonly utilized to

determine active site structure in copper enzymes, it has also

helped in elucidating mechanistic pathways. For example,

the ability of CO to bind to both PHM metal sites,

surprisingly even Cu_Hwhen peptidylglycine substrates bind,²²

at the time led to new hypotheses for probing the enzyme

mechanism and dynamics (i.e., superoxide channeling^20,22vs

hydrogen tunneling for the substrate hydrogen atom abstrac-

tion^23-27) of the uncoupled metal centers. In addition, the

course of CO and presumably O₂reactivity in the heme_a3/

Cu_Bactive site of cytochrome c oxidase (CcO) has been

probed through laser flash photolysis (LFP).^28-33Upon the

photorelease of carbon monoxide from heme_a3, CO rapidly

(ps) moves to Cu^I_Band subsequently escapes or returns to

the heme_a3on a millisecond time-scale, suggesting that Cu^I_B

can act as a doorway for substrate O₂-binding to heme_a3in

this binuclear active site of CcO. LFP is a common

methodology used to probe heme protein-actives-site envi-

ronments giving kinetic and thermodynamic insights.^33-35

Whereas CO photoejection has yet to be observed from a

protein that has only copper as the active-site metal, we

recently communicated that metal-to-ligand charge-transfer

(MLCT) excitation of at least one type of copper(I) carbonyl

complex, [Cu^I(H-tmpa)(CO)]⁺(1^H-CO; TMPA ) tris(2-

pyridylmethyl)amine); Chart 1), resulted in the efficient

release of CO.³⁶The CO rebinding was quantitative, and no

net photochemistry was observed. The rate constant for CO

rebinding following flash photolysis of 1^H-CO in tetrahy-

drofuran solvent (THF) was determined to approach a

diffusion-controlled value, k_CO) 1.9 × 10⁹M^-1s^-1at

25 °C. We were also successful at applying the flash-and-

trap technique, pioneered by Gibson and co-workers for heme

protein analysis.³⁷Light excitation (λ_ex) 355 nm) of 1^H-

CO in the presence of O₂in THF resulted in the transient

appearance of a primary copper(I)-dioxygen adduct, a

cupric-superoxo (Cu^II(O₂^•-)) species.³⁸The temperature

(10) Solomon, E. I.; Sundaram, U. M.; Machonkin, T. E. Chem. ReV. 1996,

96, 2563-2605.

(11) DellaLonga, S.; Ascone, I.; Bianconi, A.; Bonfigli, A.; Castellano, A.

C.; Zarivi, O.; Miranda, M. J. Biol. Chem. 1996, 271, 21025-21030.

(12) Matoba, Y.; Kumagai, T.; Yamamoto, A.; Yoshitsu, H.; Sugiyama,

M. J. Biol. Chem. 2006, 281, 8981-8990.

(13) Klabunde, T.; Eicken, C.; Sacchettini, J. C.; Krebs, B. Nature Struct.

Biol. 1998, 5, 1084-1090.

(14) Rompel, A.; Fischer, H.; Meiwes, D.; Buldt-Karentzopoulos, K.;

Dillinger, R.; Tuczek, F.; Witzel, H.; Krebs, B. J. Biol. Inorg. Chem.

1999, 4, 56-63.

(15) Gerdemann, C.; Eicken, C.; Krebs, B. Acc. Chem. Res. 2002, 35, 183-

191.

(16) Blackburn, N. J.; Pettingill, T. M.; Seagraves, K. S.; Shigeta, R. T. J.

Biol. Chem. 1990, 265, 15383-15386.

(17) Pettingill, T. M.; Strange, R. W.; Blackburn, N. J. J. Biol. Chem. 1991,

266, 16996-17003.

(18) Reedy, B. J.; Blackburn, N. J. J. Am. Chem. Soc. 1994, 116, 1924-

1931.

(19) Boswell, J. S.; Reedy, B. J.; Kulathila, R.; Merkler, D.; Blackburn,

N. J. Biochemistry 1996, 35, 12241-12250.

(20) Jaron, S.; Blackburn, N. J. Biochemistry 1999, 38, 15086-15096.

(21) Jaron, S.; Blackburn, N. J. Biochemistry 2001, 40, 6867-6875.

(22) Jaron, S.; Mains, R. E.; Eipper, B. A.; Blackburn, N. J. Biochemistry

2002, 41, 13274-13282.

(23) Francisco, W. A.; Knapp, M. J.; Blackburn, N. J.; Klinman, J. P. J.

Am. Chem. Soc. 2002, 124, 8194-8195.

(24) Chen, P.; Solomon, E. I. J. Am Chem. Soc. 2004, 126, 4991-5000.

(25) Bauman, A. T.; Jaron, S.; Yukl, E. T.; Burchfiel, J. R.; Blackburn, N.

J. Biochemistry 2006, 45, 11140-11150.

(26) Bauman, A. T.; Yukl, E. T.; Alkevich, K.; McCormack, A. L.;

Blackburn, N. J. J. Biol. Chem. 2006, 281, 4190-4198.

(27) Klinman, J. P. J. Biol. Chem. 2006, 281, 3013-3016.

(28) Alben, J. O.; Moh, P. P.; Fiamingo, F. G.; Altschuld, R. A. Proc.

Natl. Acad. Sci-Biol. 1981, 78, 234-237.

(29) Einarsdottir, O.; Dyer, R. B.; Lemon, D. D.; Killough, P. M.; Hubig,

S. M.; Atherton, S. J.; Lopez-Garriga, J. J.; Palmer, G.; Woodruff,

W. H. Biochemistry 1993, 32, 12013-24.

(30) Stavrakis, S.; Koutsoupakis, K.; Pinakoulaki, E.; Urbani, A.; Saraste,

M.; Varotsis, C. J. Am Chem. Soc. 2002, 124, 3814-3815.

(31) Okuno, D.; Iwase, T.; Shinzawa-Itoh, K.; Yoshikawa, S.; Kitagawa,

T. J. Am. Chem. Soc. 2003, 125, 7209-7218.

(32) McMahon, B. H.; Fabian, M.; Tomson, F.; Causgrove, T. P.; Bailey,

J. A.; Rein, F. N.; Dyer, R. B.; Palmer, G.; Gennis, R. B.; Woodruff,

W. H. Bba-Bioenergetics 2004, 1655, 321-331.

(33) Larsen, R. W.; Miksovska, J. Coord. Chem. ReV. 2007, 251, 1101-

1127.

dependence of this rate constant implied a value of k_O2

)

1.3 × 10⁹M^-1s^-1at 25 °C. Remarkably, this rate constant

for dioxygen binding was found to exceed that previously

measured for synthetic or biological hemes.^35,39-41Of even

greater significance are the nearly equal k_COand k_O2values

(34) Niederhoffer, E. C.; Timmons, J. H.; Martell, A. E. Chem. ReV. 1984,

84, 137.

(35) Momenteau, M.; Reed, C. A. Chem. ReV. 1994, 94, 659-698.

(36) Scaltrito, D. V.; Fry, H. C.; Showalter, B. M.; Thompson, D. W.;

Liang, H.-C.; Zhang, C. X.; Kretzer, R. M.; Kim, E.-i.; Toscano, J.

P.; Karlin, K. D.; Meyer, G. J. Inorg. Chem. 2001, 40, 4514-4515,

and refs cited therein.

(37) Greenwood, C.; Gibson, Q. H. J. Biol. Chem. 1967, 242, 1782-1787.

(38) Fry, H. C.; Scaltrito, D. V.; Karlin, K. D.; Meyer, G. J. J. Am Chem.

Soc. 2003, 125, 11866-11871.

(39) Collman, J. P.; Boulatov, R.; Sunderland, C. J.; Fu, L. Chem. ReV.

2004, 104, 561-588.

(40) Collman, J. P.; Brauman, J. I.; Doxsee, K. M.; Sessler, J. L.; Morris,

R. M.; Gibson, Q. H. Inorg. Chem. 1983, 22, 1427-1432.

242 Inorganic Chemistry, Vol. 47, No. 1, 2008

Copper(I) Pyridylalkylamine Compounds

Scheme 1

O₂, CO or NO) reactions with hemes and copper proteins

are discussed.

Experimental Section

Materials and Methods. All of the reagents were purchased

from commercial sources. All of the air-sensitive compounds were

synthesized utilizing Schlenk techniques. Diethyl ether (Fisher

Scientific 99+ %) and tetrahydrofuran (THF; EMD 99%) solvents

were purified by passing through a series of activated alumina

columns (Innovative Technology; Newburyport, MA). Acetonitrile

(CH₃CN, Fisher Scientific 99+ %) was dried and distilled over

calcium hydride under an argon (Air Gas East, grade 4.8)

atmosphere. The solvents CH₃CN and THF were introduced into

an MBraun Labmaster 130 dry box (O₂< 1 ppm; H₂O < 1 ppm)

after purging for 1 h with argon gas and then placed under vacuum.

Elemental analyses were performed by Desert Analytics (Tucson,

AZ) or Quantitative Technologies, Inc. (QTI; Whitehouse, NJ). ¹H

NMR spectra were recorded at 400 MHz on a Bruker Avance FT-

NMR spectrometer. Deuterated nitromethane (CD₃NO₂) was pur-

chased from Aldrich (99%) and deaerated under nitrogen gas (Air

Gas East, grade 4.8).

as has been observed in other copper proteins and heme

systems (vide infra).

Thus, CO appears as an excellent surrogate for O₂, and

further elucidation of the dynamics of CO binding to copper-

(I) complexes is of bioinorganic relevance. The ability to

probe the ligand environment and function relevant to copper

protein-actives sites is an area of great interest because it is

already established that the detailed nature of the ligand

environment around copper (i.e., donor atom type, number

of donors, coordination geometry, etc.) dictates the course

of (ligand)Cu^I/O₂chemistry, dynamics, dioxygen coordina-

tion mode, and subsequent reactivity toward substrates.^2-4

Here, we report systematic studies of CO coordination and

photodissociation for a series of copper(I) compounds

possessing tripodal tetradentate ligand analogues of TMPA.

The studies are designed to test generality and to determine

some rules for the realization of this novel photochemistry.

To probe electronic effects, the pyridyl donors of the parent

TMPA ligand have been para-substituted with electron-

donating -NMe₂or -OMe substituents (1^R, Chart 1).⁴²In

addition, the ligand chelate ring size has been systematically

increased by employing the ligand series 2^R, PMEA f

PMAP f TEPA (Chart 1).⁴³Finally, a bulkier and more

hydrophobic ligand, BQPA, contains one pyridyl and two

quinolyl arms (Chart 1).⁴⁴The choice of these ligands comes

in part from their previous use in (ligand)-copper(I)/O₂

reactivity studies.^42-44

Synthesis. Me₂N-TMPA,⁴²MeO-TMPA,⁴²H-TMPA,⁴⁵PMEA,⁴³

PMAP,⁴³BQPA,⁴⁴[Cu^I(CH₃CN)₄]B(C₆F₅)₄,⁴⁶[Cu^I(Me₂N-tmpa)]B-

(C₆F₅)₄(1^Me2N),⁴²[Cu^I(MeO-tmpa)(CH₃CN)]B(C₆F₅)₄(1^MeO),⁴²and

[Cu^I(H-tmpa)(CH₃CN)]B(C₆F₅)₄(1^H)⁴²were prepared according to

published procedures.

[Cu^I(NMe₂-TMPA)(CO)]B(C₆F₅)₄(1^Me2N-CO). In a 100 mL

Schlenk flask equipped with a stir bar, NMe₂-TMPA (34 mg, 0.081

mmol) and [Cu^I(CH₃CN)₄]B(C₆F₅)₄(71 mg, 0.078 mmol) were

added together. To this mixture, CO-saturated diethyl ether (5 mL)

was added under an atmosphere of CO. The resulting pale-yellow

solution was stirred for 75 min while CO was purged through 100

mL of heptane. The CO-saturated heptane was subsequently added

to the stirring solution, resulting in a white cloudy mixture. After

2 h of stirring under a CO atmosphere, a white solid precipitated.

Additional time was provided for the white powder to settle. The

supernatant was removed via a cannula, and the resulting white

powder was placed in vacuo for 15 min, yielding 72 mg (77%).

Anal. Calcd: C, 49.45; H, 2.79; N, 8.24. Found: C, 49.14; H, 2.66;

N, 8.14. ¹H NMR (CD₃NO₂): δ 8.68 (3H, s, br), 7.89 (3H, s, br),

7.44 (6H, s, br), 4.15 (6H, s).

Microsecond fast CO-binding rates like those reported here

can only be obtained using flash photolytic techniques; other

methods such as stopped-flow do not offer a sufficient time

resolution. Therefore, the literature involving LCu^I-CO

kinetics has been very limited and the thermodynamics (∆H^q,

∆S^q) that govern these dynamics are nonexistent (as far as

we have investigated). Here, we present the first such detailed

investigation of CO and copper(I) interactions. In addition

to new transient absorption laser flash-photolysis CO-

recombination kinetic studies, we also present here comple-

mentary chemical investigations (Scheme 1). Detailed struc-

tural, electrochemical, infrared, and equilibrium CO-binding

studies (via UV-vis spectrophotometric titrations) have been

carried out. Comparisons to data available for O₂binding to

these copper complexes as well as to small-molecule (i.e.,

[Cu^I(tmpa)(CO)]B(C₆F₅)₄(1^H-CO). In a 100 mL Schlenk flask

equipped with a stir bar, TMPA (32 mg, 0.11 mmol) and [Cu^I(CH₃-

CN)₄]B(C₆F₅)₄(100 mg, 0.11 mmol) were added together. To this,

CO saturated Et₂O (5 mL) was added under a CO atmosphere. The

resulting light-yellow solution was stirred for 30 min. CO-saturated

heptane (100 mL) was added to the solution to precipitate the

product. After stirring for 1 h, a white solid precipitated. The

supernatant was decanted, and the solid was placed in vacuo,

yielding 93 mg (80%) of a white powder. Anal. Calcd: C, 48.68;

1

H, 1.71; N, 5.28. Found: C, 48.50; H, 1.89; N, 4.49. H NMR

(CD₃NO₂): δ 8.68 (3H, s, br), 7.89 (3H, s, br), 7.44 (6H, s, br),

4.15 (6H, s).

[Cu^I(pmea)]B(C₆F₅)₄(2^pmea). [Cu^I(CH₃CN)₄]B(C₆F₅)₄(300 mg,

0.33 mmol) was placed in a 100 mL Schlenk flask equipped with

a stir. To this solid, PMEA (111 mg, 0.36 mmol) dissolved in dried

and deaerated diethyl ether was added under an argon atmosphere.

The resulting yellow solution was stirred for 30 min. Dried and

(41) Kano, K.; Kitagishi, H.; Dagallier, C.; Kodera, M.; Matsuo, T.;

Hayashi, T.; Hisaeda, Y.; Hirota, S. Inorg. Chem. 2006, 45, 4448-

4460.

(42) Zhang, C. X.; Kaderli, S.; Costas, M.; Kim, E.-i.; Neuhold, Y.-M.;

Karlin, K. D.; Zuberbu¨hler, A. D. Inorg. Chem. 2003, 42, 1807-1824.

(43) Schatz, M.; Becker, M.; Thaler, F.; Hampel, F.; Schindler, S.; Jacobson,

R. R.; Tyekla´r, Z.; Murthy, N. N.; Ghosh, P.; Chen, Q.; Zubieta, J.;

Karlin, K. D. Inorg. Chem. 2001, 40, 2312-2322.

(44) Wei, N.; Murthy, N. N.; Chen, Q.; Zubieta, J.; Karlin, K. D. Inorg.

Chem. 1994, 33, 1953-1965.

(45) Tyekla´r, Z.; Jacobson, R. R.; Wei, N.; Murthy, N. N.; Zubieta, J.;

Karlin, K. D. J. Am. Chem. Soc. 1993, 115, 2677-2689.

(46) Liang, H.-C.; Kim, E.; Incarvito, C. D.; Rheingold, A. L.; Karlin, K.

D. Inorg. Chem. 2002, 41, 2209-2212.

Inorganic Chemistry, Vol. 47, No. 1, 2008 243

Fry et al.

deaerted heptane (100 mL) was added to precipitate the complex.

After stirring for 1 h, the supernatant was decanted. The remaining

yellow-oily solid was dissolved again in dried and dearated diethyl

ether (10 mL) and precipitated with deaerated heptane (100 mL).

The resulting solid was placed in vacuo to yield 268 mg (77%) of

a yellow powder. Anal. Calcd: C, 49.33; H, 1.93; N, 5.35. Found:

C, 49.54; H, 2.21; N, 5.12. ¹H NMR (CD₃NO₂): δ 8.93 (br, 3 H),

7.81 (m, br, 3 H), 7.37 (m, br, 6 H), 4.05 (br, 4 H), 3.21 (br, 2 H),

3.01 (br, 2 H).

spectrophotometer. Low temperature measurements (-80 to 0 °C)

were made using a Neslab Ult-95 controller by circulating methanol

through a two-window optical dewar, as previously described. High-

temperature measurements (20-45 °C) were made using a Neslab

Endocal circulating (1:1 ethylene glycol/water) bath. Samples were

prepared in the dry box using a 10 mm quartz Schlenk cuvette

(designed for low-temperature studies) equipped with a rubber

serum stopper. The gas mixtures were introduced into the cuvette

using a 24 in. needle and bubbling directly through the solution

for ∼30 s at room temperature. Samples were equilibrated for a

minimum of 3 min upon changing the temperature.

Infrared Spectroscopy. IR spectra were obtained at room

temperature using a Mattson Galaxy 4030 series FTIR spectropho-

tometer. Samples were prepared in an inert atmosphere box. Isolated

copper(I) carbonyl compounds (8-10 mg) were prepared in Nujol

mull for solid-state analysis by pressing between two sodium

chloride salt plates. Solutions for the infrared spectral analysis of

carbonylated species were prepared by dissolving 8-10 mg in <1

mL of solvent (CH₃CN or THF) in a 1 dram vial capped with a

red rubber septum. The samples were brought out of the dry box,

purged with CO for ∼10 s at room temperature and reintroduced

into the dry box, where the samples were transferred to a Spectralys

Specac KBr plate solution-IR cell.

Cyclic Voltammetry. Studies were carried out using a Bioana-

lytical Systems BAS-100B electrochemistry analyzer. The air-

sensitive samples were prepared in a 10 mL Schlenk flask equipped

with a 14/20 rubber septum in an inert atmosphere box and

comprised of a 1-3 mM solution of copper compound with 100

mM tetrabutylammonium hexafluorophosphate (supporting elec-

trolyte) in CH₃CN. The flask was removed from the dry box, and

the solution was transferred under positive argon pressure via

cannula to a three-neck (25 mL) round-bottom flask equipped with

a glassy-carbon working electrode, saturated calomel reference

electrode, and a platinum-wire counter electrode. Each electrode

was connected to the flask by piercing through a 14/20 rubber

septum. An external reference of ferrocene (Fc, monitoring the Fc⁺/

Fc couple) under the same conditions described was also measured,

E_1/2) 400 mV versus SCE.

[Cu^I(pmap)]B(C₆F₅)₄(2^pmap). [Cu(CH₃CN)₄]B(C₆F₅)₄(200 mg,

0.22 mmol) was placed in a 100 mL Schlenk flask equipped with

a stir bar. To this, PMAP (70 mg, 0.22 mmol) dissolved in dried

and deaerated diethyl ether (5 mL) was added under an argon

atmosphere. The resulting yellow solution was stirred for 30 min.

Deaerted heptane (100 mL) was added to precipitate the complex.

After stirring for 1 h, the supernatant was decanted. The resulting

solid was placed in vacuo to yield 173 mg (74%) of a yellow

powder. Anal. Calcd: C, 49.81; H, 2.09; N, 5.28. Found: C, 49.90;

1

H, 2.36; N, 5.25. H NMR (CD₃NO₂): δ 8.84 (m, 3 H), 7.79 (m,

3 H), 7.35 (m, 6 H), 4.28 (s, 2 H), 3.17 (br, 4 H), 2.96 (br, 4 H).

[Cu^I(bqpa)]B(C₆F₅)₄(3^bqpa). In a 100 mL Schlenk flask

equipped with a stir bar, BQPA (29 mg, 0.074 mmol) and [Cu-

(CH₃CN)₄]B(C₆F₅)₄(65 mg, 0.071 mmol). To this, dried and

deaerated diethyl ether (5 mL) was added under an argon

atmosphere. The resulting yellow solution was stirred for 30 min.

Deaerted heptane (100 mL) was added to precipitate the complex.

After stirring for 1 h, the supernatant was decanted. The remaining

yellow-oily solid was dissolved again in dried and dearated diethyl

ether (10 mL) and precipitated with deaerated heptane (100 mL).

The resulting solid was placed in vacuo to yield 66 mg (83%) of

a yellow/orange powder. Anal. Calcd: C, 53.00; H, 1.96; N, 4.94.

1

Found: C, 52.97; H, 1.94; N, 4.95. H NMR (CD₃NO₂): δ 8.68

(3H, s, br), 7.89 (3H, s, br), 7.44 (6H, s, br), 4.15 (6H, s).

[Cu^I(bqpa)(CO)]B(C₆F₅)₄(3^bqpa-CO). In a 100 mL Schlenk

flask equipped with a stir bar, BQPA (43 mg, 0.11 mmol) and

[Cu^I(CH₃CN)₄]B(C₆F₅)₄(100 mg, 0.11 mmol) were added together.

To this, CO saturated Et₂O (5 mL) was added under a CO

atmosphere. The resulting light-yellow solution was stirred for 30

min. CO-saturated pentane (100 mL) was added to the solution

via a cannula to precipitate the product. After stirring for 1 h, a

white solid precipitated. The supernatant was decanted, and the solid

was placed in vacuo, yielding 106 mg (83%) of a white powder.

¹H NMR (CD₃NO₂): δ 8.68 (3H, s, br), 7.89 (3H, s, br), 7.44 (6H,

s, br), 4.15 (6H, s).

Gas Mixing. Carbon monoxide (Air Gas East, grade 2.3) and/

or a custom 1% CO-in-N₂gas mixture (Air Gas East) were treated

by passing through an R & D Separations oxygen/moisture trap

(Agilent Technologies OT3-4) and mixed with nitrogen (Air Gas

East, grade 4.8) through two MKS Instruments mass-flow control-

lers (MFCs, MKS 1179A) controlled by an MKS Instruments Multi

Gas Controller (MGC, MKS 647C). The gas mixtures were

determined by the set flow rates of the two gases. For example, a

10% CO mixture would be made by mixing CO at a rate of 10

standard cubic centimeters per minute (sccm) with N₂at 90 sccm

for a total flow of 100 sccm. The CO concentration in solution

was determined from the percent mixture and the literature values

for CO saturation (in THF [CO]_sat) 0.0101 M; in CH₃CN, [CO]_sat

) 0.10 M)⁴⁷and adjusted for solvent contraction.

Determination of CO Binding Constants, K_CO(M^-1). Using

spectrophotometric (UV-vis) titrations of carbon monoxide and

calculated extinction coefficients (M^-1cm^-1), K_COcould be

determined using eq 1. The molar concentration of [(L)Cu^I(CO)]⁺

was determined by eq 2 and [(L)Cu^I]⁺was determined by

subtracting [(L)Cu(CO)]⁺from the total concentration of the copper

compound being studied (eq 3). Plots of [CO] versus [(L)Cu(CO)]⁺/

[(L)Cu]⁺were found to be linear, and the slope affords the

representative equilibrium constant. Vant Hoff analysis, eq 4, of

equilibrium constants measured as a function of temperature was

used to obtain thermodynamic parameters.

[(L)Cu^I(CO)]⁺

[(L)Cu^I]⁺[CO]

k_CO

K_CO

)

(1)

(2)

k_-CO

Abs_λmax- ꢀ × [(L)Cu^I]_TOT

ꢀ_[(L)Cu(CO)]+- ꢀ_[(L)Cu]+

[(L)Cu^I(CO)]⁺)

[(L)Cu^I]⁺) [(L)Cu^I]_T⁺_OT- [(L)Cu^I(CO)]⁺

(3)

(4)

∆H° ∆S°

ln(K_CO) ) -

+

UV-Vis Spectroscopy. Absorption spectra were obtained using

a Hewlett-Packard 8452A or a Hewlett-Packard 8453 diode array

RT

R

Laser Flash Photolysis: Determination of the Bimolecular

Rate Constant, k_CO(M^-1s^-1). Transient absorbance measurements

were performed using a Nd:YAG (Continuum Surelite III) laser

(47) Cargill, R. W.; Battino, R. Carbon Monoxide, lst ed.; Pergamon

Press: New York, 1990.

244 Inorganic Chemistry, Vol. 47, No. 1, 2008

Copper(I) Pyridylalkylamine Compounds

Table 1. Numerical Crystal and Refinement Data for the X-ray Crystal Structures

[Cu^I(H-tmpa)(CO)]B(C₆F₅)₄

(1^H-CO)

[Cu^I(Me₂N-tmpa)(CO)]B(C₆F₅)₄

[Cu^I(bqpa)(CO)]B(C₆F₅)₄

complex

formula (sum)

(1^Me2N-CO)

(3^bqpa-CO)

C₄₃H₁₈B Cu F₂₀N₄O

C₅₃H₄₃B Cu F₂₀N₇O₂

1264.29

triclinic

C₅₁H₂₂B Cu F₂₀N₄O

fw

1060.96

monoclinic

P2₁/c

15.0630(9)

18.9983(9)

14.8122(11)

90

107.037(6)

90

4052.8(4)

4

1161.08

monoclinic

P2/c

21.9205(7)

7.8885(3)

26.6533(10)

90

96.523(3)

90

4579.0(3)

4

cryst syst

space group

a (Å)

b (Å)

c (Å)

P1h

11.0733(10)

15.0029(11)

16.9685(16)

69.838(8)

87.098(8)

84.613(7)

2634.1(4)

2

R (deg)

â (deg)

γ (deg)

V (Å³)

Z

µ/mm^-1(Mo KR)

reflns collected (total)

0.67587

36 957

13 927

0.52931

35 413

17 308

0.607

49 165

8356

R_int(no. of equiv. reflections)

observed reflns [I/s(I) > 2]

Final R, Rw [I/σ(I)>2]

7935

0.0418, 0.0986

9920

0.0609, 0.1911

6130

0.0393, 0.0930

on a previously described apparatus.⁴⁸Samples of copper complexes

were prepared within the dry box in a 10 mm quartz cuvette

equipped with a Schlenk stopcock and 14/20 joint. The samples

were irradiated with 355 nm pulsed light (8-10 mJ/cm²) and

protected from the probe beam with appropriate UV filters. Kinetic

data at monitored wavelengths were averages of 40 to 90 laser

pulses. The CO concentrations were varied as described above

(UV-vis spectroscopy) to determine the second-order rate constant,

X-ray Crystallographic Studies of (1^H-CO), (1^Me2N-CO), and

(3^bqpa-CO). All of the crystals reported here were prepared in a

dry box via the dissolution of 5 mg of copper complex in dry

nitrogen-saturated diethyl ether (∼0.5 mL). The yellow solutions

were transferred to a 9” NMR tube and capped with a rubber

septum. The samples were removed from the drybox, and CO was

sparged directly through the solution with a needle, followed by

the layering of CO-saturated pentane (>98%). Colorless to light-

yellow crystals were afforded after 24 h.

Crystals of each compound were placed in the N₂cold stream

at 110 K of an O.D. Xcaliber3 system equipped with a graphite

monochromator and an Enhance (Mo) X-ray Source (λ ) 0.71073

Å) operated at 2 kW power (50 kV, 40 mA). The detector was

placed at a distance of 50 mm from the crystal. The frames were

integrated with the Oxford Diffraction CrysAlisRED software

package. All of the structures were solved and refined using the

Bruker SHELXTL (ver. 6.1) software package. Analysis of the data

showed no negligible decay. Relevant crystallographic information

is given in Table 1.

k

_CO, using eq 5. Variable low-temperature studies (-80 to 0 °C)

were performed using the setup described in previous work.³⁸From

the variable temperature work, Eyring analyses, eq 6, were

performed to obtain activation parameters.

k_obs) k_CO[CO]

(5)

(6)

∆H^q∆S^q

RT

k_CO× h

In

) -

+

(

)

k_B× T

R

Comparative Actinometry. To determine quantum efficiencies,

samples of the various copper complexes were prepared in a 10

mm quartz cuvette on a millimolar scale using dried and distilled

THF in a dry box at room temperature. Upon removal from the

dry box, the samples were sparged with CO gas. The absorbance

at the excitation wavelength 355 nm was measured (for example

0.1 au) using an HP 8453 UV-vis diode array. An actinometer

(φ_eff) 1) using [Ru(bpy)₃]Cl₂in THF (with the aid of excess

NaBArF in situ metathesis) was prepared.⁴⁹The actinometer

solution was prepared to ensure that the measured absorbance at

355 nm matched that of the copper-carbonyl solution (for example

0.1 au). The samples were then irradiated at 355 nm, and a change

in absorbance (∆A) was measured at the corresponding λ_maxvalues

where the change in extinction coefficients (∆ꢀ) is known. From

these data, the concentration of transient species can be measured

using Beer’s law (∆A/∆ꢀ ) ∆Concentration). Because the quantum

yield for the formation of the [Ru(bpy)₃]²⁺MLCT excited state is

known to be 1.00, the concentration of the excited state formed is

equal to the concentration of absorbed photons. With a separate

measurement of the concentration of the copper-solvento com-

pound (CO dissociated) measured under the same conditions, the

quantum yield is determined by the ratio, that is, φ ) [(L)Cu^I]⁺/

[photons].

Results and Discussion

Synthesis of Cu(I) Carbonyl Complexes. All of the

ligands used for the present study (Chart 1) were previously

synthesized and characterized (Experimental Section),^42-44,50

as were the copper(I) complexes of the substituted TMPA

(R-TMPA) ligands,⁴²1^R(R ) Me₂N, MeO, H). Copper(I)

-

complexes as B(C₆F₅)₄salts with PMAP, PMEA, and

BQPA are described here. Copper carbonyl complexes [(L)-

Cu^I(CO)]B(C₆F₅)₄, with the exception of [Cu^I(MeO-TMPA)-

(CO)]B(C₆F₅)₄1^MeO-CO, were isolated by 1:1 mixing of

ligand and [Cu^I(CH₃CN)₄]B(C₆F₅)₄⁴⁶in CO-saturated diethyl

ether and precipitated by addition of a > 10 fold excess (by

volume) of CO saturated heptane; 1^MeO-CO was produced

in situ by sparging LCu^Icomplex solutions with CO. The

B(C₆F₅)₄^-counteranion was chosen to afford greater solubil-

ity of the resulting copper complex in a wide array of

solvents, in particular THF. Crystals suitable for X-ray

diffraction were prepared via the slow diffusion of pentane

with diethyl ether solutions of complexes under an atmo-

sphere of CO.

(48) Bignozzi, C. A.; Argazzi, R.; Chiorboli, C.; Scandola, F.; Dyer, R.

B.; Schoonover, J. R.; Meyer, T. J. Inorg. Chem. 1994, 33, 1652-

1659.

(49) Yoshimura, A.; Hoffman, M. Z.; Sun, H. J. Photochem. Photobiol. A

1993, 70, 29-33.

(50) Karlin, K. D.; Shi, J.; Hayes, J. C.; McKown, J. W.; Hutchinson, J.

P.; Zubieta, J. Inorg. Chim. Acta 1984, 91, L3-L7.

Inorganic Chemistry, Vol. 47, No. 1, 2008 245

Fry et al.

Scheme 2

Solution and Solid-State Coordination Geometries.

There is compelling evidence from the present and previous

work^45,51that one pyridyl ring of the H-TMPA ligands in

[Cu^I(H-tmpa)(CO)]⁺(1^H-CO) dissociates in solution to

accommodate an ancillary ligand such as CO. In fluid

solution, an equilibrium like that shown in Scheme 2 is found

to exist under many experimental conditions. In fact,

inductive and steric effects can be used to tune the equilib-

rium. This is particularly relevant to the present work, as

the nominally four-coordinate species, that is, with copper-

(I)-ligand tridentate chelation plus the CO donor, is identi-

fied as the species that readily photoreleases CO (see

discussion below).

X-ray Crystallography. Figure 1 shows ORTEP diagrams

of [Cu^I(H-tmpa)(CO)]⁺(1^H-CO) and [Cu^I(Me₂N-tmpa)-

(CO)]⁺(1^Me2N-CO). Partial bond length and angle data are

provided in the figure legend, with full details given in the

Supporting Information.⁵²The structures of 1^H-CO and

1^Me2N-CO display an overall five-coordinate geometry,

which has been observed previously for related analogues

Figure 1. ORTEP diagrams of A. [Cu^I(H-tmpa)(CO)]⁺(1^H-CO) and B.

[Cu^I(Me₂N-tmpa)(CO)]⁺(1^Me2N-CO). Hydrogen atoms and counteranions

are omitted for clarity. Selected bond lengths are: A. Cu-C 1.852(2), C-O

1.124(2), Cu-N_am2.479(15), and Cu-N_py(ave)2.115(15) Å; and B. Cu-C

1.870(2), C-O 1.091(3), Cu-N_am2.446(2), and Cu-N_py2.114(8) Å.

such as [Cu^I(H-tmpa)(CH₃CN)]⁺(1^H) (Cu-N_pyridyl(ave)

)

2.090(11), Cu-N_alkylamine) 2.431(10) Å; the cuprous ion lies

0.55 Å out of the N_pyplane toward the acetonitrile ligand).⁵³

Thus, the structures are more like pseudo-tetrahedral, with

the three N_pyridylligands (Cu-N_pyridyl(ave)) 2.115 (15) Å)

and the carbon from the CO bound as strong donors; the

fifth, weakly interacting apical alkylamino nitrogen is forced

out of the N_pyplane and has a bond distance of ∼2.5 Å. For

both structures (1^H-CO and 1^Me2N-CO), the average

N_pyridyl-Cu-C angle is 107° and the average N_pyridyl-Cu-

N_pyridylangle is 112°. Constraints induced by the ligand

environment lead to a preferred four-coordinate carbonyl

species (like that given in Scheme 2) when in solution (also,

see IR discussion below), as well as to the established

dynamic behavior of such cuprous species utilizing tetraden-

tate chelators.^42,54A number of overall four-coordinate

copper(I)-CO complexes have been structurally character-

ized, and they exhibit N-Cu-C bond angles of ∼120° and

N-Cu-N ∼ 95°.^55-61

A representation of the crystal structure of [Cu^I(bqpa)-

(CO)]⁺(3^bqpa-CO) is given in Figure 2, along with a listing

of selected bond lengths. The most interesting feature of this

complex is the displaced dangling quinolyl arm of the ligand

(which in fact is disordered over two positions, see Sup-

porting Information).⁵²This results in a tridentate (N₃) ligand

coordination mode in spite of the ligand’s tetradentate (N₄)

nature. The N_{pyridyl,quinolyl}-Cu-N_alkylaminoangles are quite

severe (∼80°). Thus, like other tridentate ligand copper

carbonyl complexes, the observed metal coordination ge-

ometry is distorted tetrahedral.^55-61

In previous work, the crystal structure of [Cu^I(bqpa)]B-

(C₆F₅)₄(3^bqpa) was obtained in which the pyridyl and both

quinolyl N-donors are coordinated to the cuprous ion.⁶²Note

that, in the absence of CO, 3^bqpais a four-coordinate species

(57) Karlin, K. D.; Tyekla´r, Z.; Farooq, A.; Haka, M. S.; Ghosh, P.; Cruse,

R. W.; Gultneh, Y.; Hayes, J. C.; Toscano, P. J.; Zubieta, J. Inorg.

Chem. 1992, 31, 1436-1451.

(51) Kretzer, R. M.; Ghiladi, R. A.; Lebeau, E. L.; Liang, H.-C.; Karlin,

K. D. Inorg. Chem. 2003, 42, 3016-3025.

(52) See Supporting Information.

(53) Lim, B. S.; Holm, R. H. Inorg. Chem. 1998, 37, 4898-4908.

(54) Jacobson, R. R. Ph.D. Dissertation, State University of New York at

Albany, 1989.

(55) Pasquali, M.; Marini, G.; Floriani, C.; Gaetanimanfredotti, A.; Guastini,

C. Inorg. Chem. 1980, 19, 2525-2531.

(56) Kitajima, N.; Fujisawa, K.; Fujimoto, C.; Moro-oka, Y.; Hashimoto,

S.; Kitagawa, T.; Toriumi, K.; Tasumi, K.; Nakamura, A. J. Am. Chem.

Soc. 1992, 114, 1277-1291.

(58) Ardizzoia, G. A.; Beccalli, E. M.; Lamonica, G.; Masciocchi, N.;

Moret, M. Inorg. Chem. 1992, 31, 2706-2711.

(59) Imai, S.; Fujisawa, K.; Kobayashi, T.; Shirasawa, N.; Fujii, H.;

Yoshimura, T.; Kitajima, N.; Moro-oka, Y. Inorg. Chem. 1998, 37,

3066-3070.

(60) Conry, R. R.; Ji, G. Z.; Tipton, A. A. Inorg. Chem. 1999, 38, 906-

913.

(61) Reger, D. L.; Collins, J. E. Organometallics 1996, 15, 2029-2032.

(62) Lucchese, B. Recognizing and Predicting General Trends in the

Reactivity and Chemistry of a Series of Related Copper Complexes.

Ph.D. Dissertation, Johns Hopkins University, 2003.

246 Inorganic Chemistry, Vol. 47, No. 1, 2008

Copper(I) Pyridylalkylamine Compounds

because of increased π back-bonding. However, when these

complexes are dissolved in CH₃CN, the ν_COvalues shift

significantly to higher energy, by 15-19 cm^-1, Table 2. The

loss of a pyridyl nitrogen donor is entirely consistent with

weaker π back-bonding resulting in a stronger CO bond and

thus a higher ν_COvalue. By contrast, in THF solvent, two

IR bands are observed for 1^Me2N-CO, which reasonably

correspond to the ν_COvalues for the individual pure isomers

(four-coordinate and five-coordinate), thus an equilibrium

mixture exists (Scheme 2). In fact, the Cu(I)-carbonyl

complex of MeO-TMPA displays similar behavior, except

that a mixture also exists in the solid state; two ν_COvalues

are observed in the IR spectra,⁵²corresponding to each isomer

form (Table 2).

Additional evidence for the conclusion presented here

come from a previous study where we investigated the IR

spectrum of [Cu^I(PY1)(CO)]⁺{PY1 ) the tridentate ligand

bis-(2-picolyl)amine; Chart 2}.⁵¹Notably, ν_CO) 2092 cm^-1

in CH₃CN for this complex exactly corresponds to that found

for 1^H-CO (Table 2). Thus, in solution the R-TMPA

copper carbonyl complexes 1^R-CO possess displaced py-

ridyl arms, and the cuprous ion is in an N₃coordination

environment that does not resemble their respective crystal

structures.

Unlike the R-TMPA complexes 1^R-CO, solution (THF

and CH₃CN) and solid-state IR spectra for 3^bqpa-CO indicate

very similar ν_COvalues (2091-2095 cm^-1, Table 2); thus

solution and solid-state structures must be like that observed

in the X-ray structure (vide supra) of 3^bqpa-CO, with BQPA

behaving as a tridentate ligand possessing a dangling quinolyl

arm.

With ligands where additional methylene groups were

placed between the tertiary alkylamine and the pyridyl arms,

as in complex 2^pmea-CO and 2^pmap-CO (Chart 1), a single

carbonyl stretching frequency (ν_CO) 2091_avecm^-1) is

observed in both CH₃CN and THF solvents (Table 2). This

is consistent with N₃ligand binding in solution. Note that

solid-state carbonyl complexes could not be isolated because

of the relatively low affinity of the cuprous compound

precursors 2^Rfor CO, corroborated by K_COmeasurements

(vide infra). In fact, in [Cu^I(tepa)]⁺(tepa ) tris(2-pyridyl-

ethyl)amine), with all of ligand arms expanded to -CH₂-

CH₂- moieties, there was no evidence for CO coordination

to copper in solution or in the solid state.

Figure 2. ORTEP diagram of [Cu^I(bqpa)(CO)]⁺(3^bqpa-CO). Hydrogen

atoms and counteranions are omitted for clarity. Selected bond lengths:

Cu-C 1.793(3), C-O 1.132(4), Cu-N_am2.140(2), and Cu-N_py2.036(2)

Å.

and does not bind a solvent molecule (i.e., CH₃CN);

similarly, 1^Me2N, 2^pmea, and 2^pmaphave related static

structures.^42-44,62Upon the addition of PPh₃to 3^bqpa, a

quinolyl arm dissociates, resulting in a pseudo-tetrahedral

complex;⁴⁴a similar structure with a dangling pyridyl arm

was observed in [Cu^I(H-tmpa)(PPh₃)]⁺.⁴⁵Thus, the CO

adduct here exhibits a similar structure. We suggest that the

quinolyl arm is preferentially displaced by CO because of

its greater steric bulk and demands compared to a pyridyl

ligand arm. The observation of a dangling quinolyl arm in

the solid-state structure of 3^bpqa-CO plays a key role in

deciphering infrared spectroscopic data (vide infra).

Infrared Spectroscopy. Evidence for equilibria like that

shown in Scheme 2 was abstracted from solution (THF and

CH₃CN) and solid state (Nujol mull) IR data obtained from

the series of copper carbonyl compounds of 1-3 (Table 2).

Carbon monoxide vibrational frequencies for synthetic (L)-

Cu^I(CO) complexes^55-61,63-70and copper carbonmonoxy

proteins^7-9,16-22fall in the range ν_CO) 2040-2125 cm^-1;

in fact, protein Cu^I-CO adducts typically exhibit ν_COvalues

20-40 cm^-1lower than model species.⁶⁹

As described above, [Cu^I(Me₂N-tmpa)(CO)]⁺(1^Me2N-CO)

and [Cu^I(H-tmpa)(CO)]⁺(1^H-CO) exhibit solid-state tet-

rahedral-like structures with all of the pyridyl nitrogens

coordinated. Corresponding Nujol mull IR spectra reveal

single ν_COvalues of 2055 and 2077 cm^-1, respectively. A

clear substituent effect is observed with the more-electron-

releasing Me₂N-TMPA ligand, substantially lowering ν_CO

In summary, carbon monoxide acts as a very good ligand

for these tripodal tetradentate ligand-copper(I) complexes

and leads to a number of structural possibilities, overall five-

coordinate and/or four-coordinate (with dangling ligand arm)

structures, as deduced by a combination of X-ray diffraction

studies and IR spectroscopy. This ability to distinguish

between four- and five-coordinate copper systems will aid

in determining the ligand environment responsible for CO

photodissociation (vide infra).

Electrochemical Studies. Cyclic voltammetric measure-

ments were carried out on 100 mM CH₃CN solutions of

copper(I) complexes 1-3 (Chart 1). The copper(II/I) redox

chemistry is classified as quasi-reversible because the anodic

and cathodic currents were approximately equal and the peak-

(63) Sorrell, T. N.; Jameson, D. L. J. Am. Chem. Soc. 1983, 105, 6013-

6018.

(64) Sorrell, T. N.; Malachowski, M. R. Inorg. Chem. 1983, 22, 1883-

1887.

(65) Gagne, R. R.; Allison, J. L.; Gall, R. S.; Koval, C. A. J. Am. Chem.

Soc. 1977, 99, 7170-7178.

(66) Gagne, R. R.; Koval, C. A.; Smith, T. J.; Cimolino, M. C. J. Am.

Chem. Soc. 1979, 101, 4571-4580.

(67) Nelson, S. M.; Lavery, A.; Drew, M. G. B. J. Chem. Soc., Dalton

Trans. 1986, 911-920.

(68) Schindler, S.; Szalda, D. J.; Creutz, C. Inorg. Chem. 1992, 31, 2255-

2264.

(69) Rondelez, Y.; Se´neque, O.; Rager, M.-N.; Duprat, A. F.; Reinaud, O.

Chem.sEur. J. 2000, 6, 4218-4226.

(70) Hernandez, R. M.; Aiken, L.; Baker, P. K.; Kalaji, M. J. Electroanal.

Chem. 2002, 520, 53-63.

Inorganic Chemistry, Vol. 47, No. 1, 2008 247

Fry et al.

Table 2. Carbonyl Stretching Frequencies for [(L)Cu^I(CO)]⁺Complexes and Cyclic Voltammetry Data for Cu(I) Complexes

compound

Nujol mull (cm^-1

)

CH₃CN (cm^-1

)

THF (cm^-1

)

E_1/2^c(mV vs Fc⁺/Fc)

[Cu^I(Me₂N-tmpa)(CO)]⁺(1^Me2N-CO)

[Cu^I(MeO-tmpa)(CO)]⁺(1^MeO-CO)

[Cu^I(H-tmpa)(CO)]⁺(1^H-CO)

[Cu^I(pmea)(CO)]⁺(2^pmea-CO)

[Cu^I(pmap)(CO)]⁺(2^pmap-CO)

[Cu^I(bqpa)(CO)]⁺(3^bqpa-CO)

2055

2067,2091

2074^a

2087^a

2092^a

2093

2090

2093

2049,2077

2063,2085

2090

2091

2089

-700

-490

-400

-360

-250

-210

2077

b

2095

2091

^aPreviously measured.^{42 b}Solid-state samples could not be isolated. ^cMeasurements on copper(I) complexes 1-3 (not carbonyl) in CH₃CN under a N₂

atmosphere.

(M^-1) in CH₃CN and THF. The copper(I) complexes display

Chart 2

a broad absorption band between 300 and 360 nm in both

solvents, this absorption is assigned as a metal-to-ligand

charge-transfer (MLCT) transition on the basis of the

intensity and energy. These MLCT absorptions undergo a

significant blue-shift upon the addition of CO, consistent with

the positive shift of the Cu(I) f Cu(II) oxidation observed

by cyclic voltammetry. Linear plots of [CO] versus [(L)Cu^I-

(CO)]⁺/[(L)Cu^I]⁺afford slopes equal to K_CO(eq 2, Experi-

mental Section and Figure 3). The equilibrium constants were

also measured as a function of temperature from 233 to 313

K. Thermodynamic parameters (Table 3) were then ab-

stracted from Van’t Hoff analyses of the data (Experimental

Section, eqs 1-4; and Supporting Information).⁵²

to-peak separation was greater than 60 mV (Table 2).

Comparisons to a ferrocene (Fc) and Fc/Fc⁺internal standard

redox couple showed that the voltammograms represent

single electron-transfer processes. In fact and not unexpect-

edly, a nice correlation exists between the copper(II/I)

reduction potentials (E_1/2, Table 2) and the ν_COfrequencies.

The complexes with more-electron-donating ligand possess

more-negative E_1/2values while displaying lower-energy

carbon monoxide stretches. This is straightforwardly ex-

plained, because a complex with greater E_1/2will have a

ligand with poor donating ability (favoring a lower oxidation

state) and thus ν_COwill decrease.

In separate experiments, the carbonyl compounds 1-CO

through 3-CO only exhibit irreversible oxidations at more-

positive potentials than seen in the absence of CO (data not

shown). This is in contrast to the case for many copper(I)-

carbonyl species where the increasing [CO] a reversible

electrochemical wave shifts to more-positive E_1/2values; such

data has been used to calculate the equilibrium CO binding

constants.^65,66,68,71Subtractively normalized interfacial FTIR

spectroscopy (SNIFTIRS) experiments conducted by Her-

nandez and co-workers⁷⁰indicated that the oxidation of

LCu^I-CO to Cu(II) results in the loss of the carbonyl signal

from the complex; however, the reduction of Cu(II) to Cu-

(I) under CO results in the reappearance of the LCu^I-CO

complex.

The Me₂N- versus MeO- versus H- substituted TMPA

copper(I) complexes (1^R) display significant variation in CO-

binding properties attributable to the difference in electron-

donating ability of the ligand 4-pyridyl substituents. Binding

studies in CH₃CN reveal that [Cu^I(Me₂N-tmpa)(CH₃CN)]⁺

(1^Me2N), which has the most-negative Cu^II/Iredox potential

(Table 2), coordinates CO stronger than all of the compounds

investigated (K_CO≈ 8000 M^-1, Table 3).⁵²In fact, K_COfor

1^Me2Nis so large that a custom 1% CO-in-N₂gas mixture

was necessary to quantify thermodynamic parameters (∆H°

) -40.8 kJ mol^-1, ∆S° ) -64.6 J mol^-1K^-1) within the

temperature range of 283-313 K. Weaker donor ligands such

as in 1^MeOand 1^Hhave significantly diminished equilibrium

binding constants (K_CO) 780 and 220 M^-1, respectively,

Table 3). For 1^Me2Nand 1^MeO, the ∆H° values are consider-

ably more favorable (i.e., more negative) than for 1^H,

accounting for the large differences in K_COvalues. However,

as revealed by the thermodynamic parameters listed in Table

3, ∆H° values do not vary systematically, with 1^MeO(∆H°

) -48.7 kJ mol^-1) being more negative than 1^Me2N; the

entropy changes compensate such that K_COvalues for the

series decrease monotonically, 1^Me2N> 1^MeO> 1^H.

For [Cu^I(pmea)]⁺(2^pmea) and [Cu^I(pmap)]⁺(2^pmap), where

the chelate ring size increases relative to TMPA, rather-low

K_COvalues (Table 3) were observed, 150 and 2.1 M^-1,

respectively. The K_COvalues nevertheless align with their

observed redox potentials and ν_COvalues (Table 2); PMAP

is a better donor to the cuprous ion, having high E_1/2and

ν_COvalues. The K_COand thermodynamic parameters for

2^pmea, 2^pmap, and 1^Hare roughly in the same range. We

suggest that the systematic decrease in K_COon going form

1^Hto 2^pmeato 2^pmap(220 to 2.1 M^-1) may reflect increased

steric demand for CO binding to the copper(I) center; with

Equilibrium CO Binding (K_CO) in CH₃CN. Carbon

monoxide binds to the copper(I) complexes under study to

varying degrees. In fact, there is surprisingly limited literature

on the quantitative study of CO interactions with copper(I)

complexes. The techniques that have been employed to

measure equilibrium binding constants include spectropho-

tometric,⁶⁵electrochemical,^66,68,71and manometric^63,64,67meth-

ods.

Spectrophotometric titrations of carbon monoxide into

submillimolar concentrations of copper(I) compounds 1-3

were used to determine equilibrium binding constants, K_CO

(71) Nanda, K. K.; Addison, A. W.; Paterson, N.; Sinn, E.; Thompson, L.

K.; Sakaguchi, U. Inorg. Chem. 1998, 37, 1028-1036.

248 Inorganic Chemistry, Vol. 47, No. 1, 2008

Copper(I) Pyridylalkylamine Compounds

Figure 3. Spectrophotometric titration of CO to [Cu^I(H-tmpa)(solvent)]⁺(1^H) in (a) CH₃CN and (b) THF at room temperature. (a) Black, [CO] ) 0.0 mM;

red, [CO] ) 1.0 mM; blue, [CO] ) 3.0 mM; green, [CO] ) 5.0 mM; magenta, [CO] ) 7.0 mM; thin black, [CO] ) 10 mM. The inset represents the linear

plots of [Cu^I(H-tmpa)(CO)]⁺/[Cu^I(H-tmpa)(CH₃CN)]⁺(i.e., [1^H-CO]/[1^H]) versus [CO] at various temperatures in CH₃CN solvent: Squares, 253 K; diamonds,

263 K; triangles, 273 K; circles, 293 K. b) Black, [CO] ) 0.0 mM; red, [CO] ) 0.1 mM; blue, [CO] ) 0.2 mM; green, [CO] ) 0.3 mM; magenta, [CO]

) 0.4 mM; thin black, [CO] ) 10 mM. The inset represents the linear plots of [Cu^I(H-tmpa)(CO)]⁺/[Cu^I(H-tmpa)(thf)]⁺(i.e., [1^H-CO]/[1^H]) versus [CO]

at various temperatures in THF solvent: Squares, 293 K; diamonds, 298 K; triangles, 303 K; circles, 308 K; open squares, 313 K.

longer chelate arms, the copper(I) is more fully surrounded

by the tetradentate ligand framework.

These differences in thermodynamic parameters suggest that

solvent ligation is critical and leads to differing behavior for

CO binding to 3^bqpacompared to 1^H, Scheme 3. Acetonitrile

does not coordinate to Cu(I) in 3^bqpa(vide supra), leading to

a highly negative entropy for the formation of 3^bqpa-CO.

In contrast, the loss of a CH₃CN ligand must occur prior to

CO coordination to 1^H.

Carbon Monoxide Binding in THF; Solvent Effects.

Carbon monoxide binding to the copper(I) complexes under

study was also examined in THF solvent, which is a very

weak ligand (if at all) for cuprous ion. In fact, there is a

greater than 500-fold increase in K_COfor [Cu^I(tmpa)(Solv)]⁺

(1^H) in THF versus CH₃CN. The K_COvalues could not be

obtained for the more strongly binding ligand complexes

Carbon monoxide binding to [Cu^I(bqpa)]⁺(3^bqpa) was

expected to be poor, because of the steric constraints of the

large quinolyl moieties. On the contrary, CO binds somewhat

more strongly to 3^bqpa, K_CO) 660 M^-1, than to 1^H, K_CO

)

220 M^-1(Table 3), in CH₃CN. This observation contrasts

with the relationship between redox potentials and CO

binding observed for the other copper pyridylalkylamine

complexes (vide supra). Van’t Hoff analysis indicates a much

more favorable standard enthalpy of the CO reaction with

3^bqpa(∆H° ) -36.7 compared to -25.7 kJ mol^-1for 1^H)

while accompanied by a more negative entropy change (∆S°

) -72.2 compared to -41.3 J mol^-1K^-1for 1^H) (Table 3).

Inorganic Chemistry, Vol. 47, No. 1, 2008 249

Fry et al.

Table 3. Kinetic and Thermodynamic Parameters for Various Copper Complexes (Constants Reported at 298 K)^a

compound in CH₃CN

K_CO

∆H°

∆S°

k_CO

∆H^q

b

∆S^q

b

k_-CO

∆H^q

∆S^q

[Cu^I(Me₂N-tmpa)(CO)]⁺(1^Me2N-CO)

[Cu^I(MeO-tmpa)(CO)]⁺(1^MeO-CO)

[Cu^I(H-tmpa)(CO)]⁺(1^H-CO)

[Cu^I(pmea)(CO)]⁺(2^pmea-CO)

[Cu^I(pmap)(CO)]⁺(2^pmap-CO)

[Cu^I(bqpa)(CO)]⁺(3^bqpa-CO)

8000

780

220

150

2.1

-40.8

-48.7

-25.7

-32.8

-21.9

-36.7

-64.6

-108

-41.3

-68.2

-45.4

-72.2

1.6 × 10⁸

1.79 × 10⁴

n/a

60.7

n/a

50.6

4.8 × 10⁷

11.7

-58.6

6.28 × 10⁴

5.9 × 10⁷

17.9

-36.1

2.68 × 10⁵

43.6

5.2

c

660

4.6 × 10⁶

12.9

-74.1

6.97 × 10³

49.6

-1.9

compound in THF

K_CO

∆H°

∆S°

k_CO

∆H^q

∆S^q

k_-CO

∆H^qd

∆S^qd

[Cu^I(H-tmpa)(CO)]⁺(1^H-CO)

[Cu^I(pmea)(CO)]⁺(2^pmea-CO)

[Cu^I(pmap)(CO)]⁺(2^pmap-CO)

[Cu^I(bqpa)(CO)]⁺(3^bqpa-CO)

1.25 × 10⁵

-35.9

-38.1

-28.4

-44.7

-42.6

-86.9

-80.4

-84.2

1.92 × 10⁹

7.13

-43.3

1.54 × 10⁴

43.0

-0.7

c

140

c

6.0

2730

3.00 × 10⁷

5.35

-83.8

1.10 × 10⁴

50.05

0.4

^aUnits of kinetic and thermodynamic values are as follows: K_CO, M^-1; k_CO, M^-1s^-1; k_-CO, s^-1; ∆H, kJ mol^-1; ∆S, J mol^-1K^{-1 b}Data analysis

.

complicated by excited-state formation and the inability to confidently obtain k_CO. See Discussion for details. ^cNo evidence for CO photodissociation was

obtained. ^dSee Supporting Information for Van’t Hoff plots. ^en/a ) not available.

Scheme 3

there is no evidence for solvent coordination (i.e., of solvent

molecules) beyond the N₄donation by the tetradentate

ligands. As previously mentioned, the extended ligand arms

and resulting increased chelate ring satisfies the coordination

needs of the cuprous ion.

Summary of CO-Binding Eqilibrium Data. Through

spectrophotometic titrations, quantitative insights into CO

binding to copper(I) coordination complexes have been

obtained. As expected from basic coordination chemistry

principles, the electron-donating ability of the ligand as well

as the geometry have an effect on the equilibrium constants.

Increasing the electron-donating ability of the ligand, such

as in 1^R, leads to enhanced CO-binding affinities; however,

increasing the chelate ring size of the ligand, such as in 2^R,

leads to a lower affinity for CO. The effect of solvent (CH₃-

CN) ligation is significant for 1^Hand results in differing CO

binding behavior for 3^bqpa

.

1^MeOand 1^Me2Nbecause the CO binding affinity was too

great; the equilibrium lies too far to the right, that is toward

[(L)Cu^I(CO)]⁺. Even with the use of a 1% CO-in-N₂gas

mixture, measurable differences in concentration were not

observed. The origin of the enhanced binding of CO in THF

is enthalpic (Table 3).

In contrast to the behavior observed in CH₃CN, CO binds

more favorably to [Cu^I(tmpa)(thf)]⁺(1^H) than [Cu^I(bqpa)]⁺

(3^bqpa) in THF solvent by greater than four orders of

magnitude (Table 3). The stronger CO binding in THF now

follows expectations in terms of observed copper(I) complex

redox potentials (E_1/2) and ν_COvalues, that is, TMPA is

inherently a better donor ligand to copper. Acetonitrile

binding and the lack thereof for 3^bqpathus caused the switch

in affinities. Notice that the enthalpic term for the Cu(I)-

CO interaction is much more favorable when solvent binding

is not a factor; in THF versus CH₃CN, ∆H° decreases by

10.2 kJ mol^-1for CO binding to 1^Hand by 8.0 kJ mol^-1for

3^bqpa. Further examination of Table 3 shows that the reaction

entropies for 1^Hand 3^bqpaare nearly independent of the

solvent.

It is of interest to note that our equilibrium constants are

consistent with those previously measured for other copper

complexes. Nelson and co-workers measured CO binding

to a series of copper(I) coordination complexes with acyclic

Schiff-base ligands possessing three nitrogen donors, K_CO

) 1.6-32.1 M^-1.⁶⁷The bulkiness of the ligands lead to weak

CO interactions and are thus similar to what we observe for

3^bqpa. In a complex that is sterically unhindered, Gagne´ and

co-workers used an N₄macrocycle, LBF₂(LBF₂) 1,1-

difluoro-4,5,11,12-tetramethyl-1-bora-3,6,10,13-tetraaza-2,-

14-dioxacyclotetradeca-3,5,10,12-tetraenate), to examine CO

binding properties.⁶⁵The copper(I) ion in [Cu(LBF₂)(CO)]

rests a full 0.96 Å out of the N₄plane, such that, when CO

binds, an unusual five-coordinate copper-carbonyl complex

forms. The out-of-the-plane coordination results in a large

binding constant (K_CO) 4.7 × 10⁴M^-1) similar to what we

observe in our R-TMPA copper complexes, 1^R.

Enhanced CO binding affinities have also been observed

in binuclear dicopper ligand systems. Gagne´ et al.⁶⁶measured

the equilibrium constant for a single CO molecule binding

to a mixed-valence Cu(II)-Cu(I) complex as K_CO) 2.8 ×

10⁴M^-1, and Schindler et al.⁶⁸measured the binding of two

CO molecules per binuclear copper(I) complex as 1.4 × 10⁴

M^-1per Cu(I). The latter binucleating ligand system {1,3-

bis[bis(2-pyridylmethyl)amino]benzene} contains the same

Examination of the K_COdata obtained for [Cu^I(pmea)]⁺

(2^pmea) and [Cu^I(pmap)]⁺(2^pmap) (Table 3) indicates that the

solvent has little effect on equilibrium binding constants and

thermodynamic parameters. This is consistent with the

chemistry previously observed for these complexes,⁴³in that

250 Inorganic Chemistry, Vol. 47, No. 1, 2008

Copper(I) Pyridylalkylamine Compounds

six-membered chelate moiety, as found in 2^R. The higher

binding affinity in Schindler’s system is most likely due to

the absence of a third pyridyl donor per copper, making for

a more coordinatively unsaturated complex. In a rigid

binuclear copper(I) system, Nanda et al. concluded that ligand

flexibility has a more effective influence on CO-adduct

formation than increased donor basicity.⁷¹

Such behavior has never before been observed for copper

complexes (1^H-CO was the first reported example).³⁶

Previous studies photolyzing copper(I) carbonyl com-

plexes^72,73and/or carbonmonoxy copper proteins^74-76resulted

in luminescence and no observed CO dissociation from the

metal center.

For the R-TMPA copper complexes 1^R, it was expected

that the electron-donating ability of the ligand would

influence the CO-binding rate constant (k_CO) as follows:

1^Me2N-CO > 1^MeO-CO > 1^H-CO. Complying with this

trend, CO binding to 1^Me2Nin CH₃CN indeed possesses the

greatest rate constant (k_CO(298 K) ) 1.6 × 10⁸M^-1s^-1,

Table 3). Activation parameters for 1^Me2Nwere not obtained

because of competing photophysical events observed at lower

temperatures; details of such aspects will be the subject of a

future publication. The rate constants for CO rebinding to

1^MeOand 1^Hin CH₃CN are similar with k_CO(298 K) ) 4.8

× 10⁷and 5.9 × 10⁷M^-1s^-1, respectively. It was

hypothesized that the rate constant for 1^MeOwould be larger

than that for 1^Hbecause of the more-electron-rich copper

center. The unexpected trend is attributed to CH₃CN binding

more favorably to 1^MeOthan 1^H, which competitively

decreases the rate at which CO binds to copper. Eyring

analysis indicates a favorable/lower enthalpy of activation

for 1^MeO, but the overall rate constant was compromised by

the more-negative activation entropy (Table 3). This reaction

entropy value may comment on the extent to which the

pyridyl arm of the ligand is coordinated (N₄) versus un-

coordinated (N₃) in the transition state (Scheme 4). However,

these values are relative to the coordination environment (N₃

vs N₄) of the starting complexes, making it difficult to

compare the parameters of one system to another.

The rate of CO recombination to [Cu^I(bqpa)]⁺(3^bqpa) in

CH₃CN is k_CO(298 K) ) 4.6 × 10⁶M^-1s^-1. The activation

parameters (∆H^q, 12.9 kJ mol^-1; ∆S^q, -74.1 J mol^-1K^-1)

are both lower than those obtained for CO binding to 1^H,

resulting in a smaller rate constant. We suggest that the bulky

quinolyl arms of the BQPA ligand sterically hinder CH₃CN

from strongly coordinating to 3^bqpa(as stated previously),

therefore the lower activation energy for CO rebinding to

3^bqpain comparison to 1^His due to diminished CO/CH₃CN

competition.

Effects of Solvent Upon CO Recombination (k_CO). In

THF solvent, CO photodissociation was spectroscopically

observable, and analyses could be carried out for 1^H-CO

and 3^bqpa-CO. The quantum yields (φ) for CO photodis-

sociation were determined by comparative actinometry: 1^H-

CO, φ )1.00 ( 0.02 and 3^bqpa-CO, φ ) 0.30 ( 0.02. Under

the same conditions in THF, photoexcitation of 1^Me2N-CO

and 1^MeO-CO exhibits absorption features in the visible

region, suggestive of a MLCT excited-state, in addition to

the expected CO rebinding; as stated previously, this will

Sorrell and co-workers investigated CO binding to coor-

dinatively unsaturated copper(I) complexes containing two-

and three-coordinate ligands, resulting in extremely low

binding affinities.^63,64Among the series of two-coordinate

+

copper(I) species with unidentate N-donor ligands, CuL₂,

the bis(imidazole) analogues do not readily react with CO.⁶³

The natural assumption is that this is due to steric constraints

from complexation with additional imidazolyl nitrogen-

+

donors forming CuL₄; however, the addition of excess

ligand such as 4-methylimidazole (4-MeIm) to [Cu(4-

MeIm)₂]⁺promotes CO coordination. Therefore, the inertness

of the linear two-coordinate copper(I) complexes is due to

their electronic structure and greater orbital mixing, leading

to shorter Cu-N bonds (∼1.87 Å). In a separate study, the

low equilibrium constants for tridentate pyrazolyl copper(I)

complexes binding CO are attributed to greater stability of

the decarbonylated form.⁶⁴Similar to the two-coordinate Cu-

(I) species, stabilization requires greater overlap between the

bonding orbitals on copper and those of the ligand pyrazole

nitrogens.

CO Photodissociation and Thermal Coordination Ki-

netics in CH₃CN (k_CO). As indicated in the introductory

material, a new aspect of the present work is to accompany

equilibrium measurements with kinetic parameters in Cu-

(I)-CO reactions. Kinetic analysis of CO coordination to

copper was accomplished through laser flash photolysis

experiments carried out between 193 and 318 K. Nanosecond

absorption spectroscopic evidence shows that CO photodis-

sociation occurs upon single-wavelength 355 nm excitation

of the copper carbonyl compounds [Cu^I(Me₂N-tmpa)(CO)]⁺

(1^Me2N-CO), [Cu^I(MeO-tmpa)(CO)]⁺(1^MeO-CO), [Cu^I-

(tmpa)(CO)]⁺(1^H-CO),³⁶and [Cu^I(bqpa)(CO)]⁺(3^bqpa

-

CO). Following photoinitiated loss of CO, these compounds

change from nearly colorless solutions to weakly yellow, as

indicated by the positive absorption changes with maxima

in the range of 320-360 nm that tail into the visible region.

The differences in extinction coefficients (∆ꢀ) range from

3000 to 7000 M^-1cm^-1. Subsequent to photodissociation,

CO rebinds to the cuprous ion, making for a fully reversible

process (Scheme 1). Simulated absorption difference spectra

calculated from the solvento (or naked) species minus the

carbonyl species are in excellent agreement with the observed

transient data, that is, Abs[Cu^I(tmpa)(solvent)]⁺- Abs[Cu^I-

(tmpa)(CO)]⁺. Thus, for these copper-carbonyl complexes,

photolysis results in CO dissociation and formation of the

otherwise isolable solvated or unsolvated [(L)Cu^I]⁺com-

plexes, followed by CO rebinding. Representative data

observed after pulsed laser excitation of 1^H-CO is shown

in Figure 4; corresponding information for the other com-

plexes is given in the Supporting Information.⁵²

(72) Sorrell, T. N.; Borovik, A. S. Inorg. Chem. 1987, 26, 1957-1964.

(73) Sorrell, T. N.; Borovik, A. S. J. Am. Chem. Soc. 1987, 109, 4255-

4260.

(74) Kuiper, H. A.; Agro, A. F.; Antonini, E.; Brunori, M. Proc. Natl. Acad.

Sci-Biol 1980, 77, 2387-2389.

(75) Finazziagro, A.; Zolla, L.; Flamigni, L.; Kuiper, H. A.; Brunori, M.

Biochemistry 1982, 21, 415-418.

Inorganic Chemistry, Vol. 47, No. 1, 2008 251

Fry et al.

Figure 4. Transient absorption difference spectra observed after 355 nm light excitation of [Cu^I(H-tmpa)(CO)]⁺(1^H-CO) in CH₃CN (top) and THF

(bottom) at room-temperature that result in CO dissociation and the formation of [Cu^I(H-tmpa)(solvent)]⁺(1^H) and subsequent observable recombination

shown at various delay times. In CH₃CN (top) squares, 0 µs; circles, 1 µs; triangles, 2 µs; diamonds, 3 µs; and stars, 10 µs. A calculated difference spectrum

of [Cu^I(H-tmpa)(CH₃CN)]⁺- [Cu^I(H-tmpa)(CO)]⁺in CH₃CN is shown in red. A kinetic trace (top inset) following CO recombination to copper at 350 nm

with a first-order fit, k_obs) 5.5 × 10⁵s^-1, shown in red. In THF (bottom) squares, 0 ns; circles, 50 ns; triangles, 100 ns; diamonds, 150 ns; and stars, 200

ns. A calculated difference spectrum of [Cu^I(H-tmpa)(THF)]⁺- [Cu^I(H-tmpa)(CO)]⁺in THF is shown in red. A kinetic trace (bottom inset) following CO

recombination to copper at 330 nm with a first-order fit, k_obs) 1.6 × 10⁶s^-1, shown in red.

be discussed elsewhere. Spectroscopic evidence for CO

photodissociation was not observed for [Cu^I(pmea)(CO)]⁺

(2^pmea-CO) nor [Cu^I(pmap)(CO)]⁺(2^pmap-CO) in either

solvent.

Acetonitrile acts as a much better ligand to Cu(I) than

THF. Therefore, CO rebinds to copper faster in the weakly

coordinating noncompeting solvent THF. When changing

from CH₃CN to THF, the rate constant (k_COat 298 K) for

1^Hincreased by greater than 3 orders of magnitude, whereas

that for 3^bqpaincreased by slightly less than 1 order of

magnitude (Table 3).⁵²The coordination of CH₃CN is

strongly implied in the enthalpy of the activation term for

1^Hand 3^bqpa. When changing from CH₃CN to THF in 1^H,

∆H^qsignificantly decreases by 10.8 kJ/mol. In the more

sterically hindered case, 3^bqpa, the trend continues with ∆H^q

decreasing by 7.6 kJ/mol. Upon changing from coordinating

to non-coordinating solvent, the decreased enthalpic barrier

indicates that the de-ligation of CH₃CN precedes CO

rebinding. The solvent change has very small affects on the

∆S^qterms, suggesting that the greatest contributor to the

overall rate of recombination is the bond-formation process.

The nearly solvent-independent activation entropies (∆S^q)

are very close in value to the actual overall reaction entropies

(∆S°) (Table 3). This similarity indicates a transition state

(76) Beltramini, M.; Dimuro, P.; Rocco, G. P.; Salvato, B. Arch. Biochem.

Biophys. 1994, 313, 318-327.

252 Inorganic Chemistry, Vol. 47, No. 1, 2008

Copper(I) Pyridylalkylamine Compounds

Scheme 4

that closely resembles the final carbonylated product (i.e., a

late transition state, Scheme 4). Even though the solvent

interacts with the metal center, the reaction observed simply

follows two reactants (copper complex and CO), forming

one product (copper terminal carbonyl complex), which as

expected yields a negative activation entropy. Negative

entropic values are often identified as associative-type

mechanisms. However, it would be premature to assign

associative or dissociative mechanisms to CO coordination

because the observed bimolecular rate constants are com-

plicated by competitive solvent interactions with the cuprous

center, which induces dissociation of ligand pyridyl arms.

Yet, changing from coordinating (CH₃CN) to weakly coor-

dinating (THF) solvent alleviates the competition between

the solvent and CO, as indicated by the large drop in the

∆H^qterms for 1^Hand 3^bqpa. We suggest that if an even-

less-coordinating solvent than THF was employed, ∆H^qmay

continue to decrease and the rate constant (k_CO) may continue

to increase.

3^bqpa-CO has the smallest off-rate of all of the compounds

studied in CH₃CN, k_-CO) 6.97 × 10³s^-1, ∆H^q, 49.6 kJ

mol^-1. We presume the low off-rate is due to the steric

hindrance of the bulky BQPA ligand, as the overall K_COfor

3^bqpa-CO formation is lower than the R-TMPA complexes

1^H-CO and 1^MeO-CO. The greatest contribution to the

overall activation energy comes from the large ∆H^qterm

obtained for both 1^H-CO and 3^bqpa-CO. The ∆S^qterm for

these compounds is virtually 0 J mol^-1K^-1, suggesting a

very small overall change from the coordinated carbonyl

complex to the transition state (i.e., an early transition state,

Scheme 4). In sharp contrast to the forward reaction (k_CO),

∆H^qfor the reverse reaction (k_-CO) shows no dependence

on solvent. The relatively large ∆H^qvalue corresponds

directly to the simple weakening and breaking of the Cu-

CO bond. Thus, the transition state for the simple dissociation

of CO from [(L)Cu^I(CO)]⁺complexes resembles the terminal

carbonyl complex (the converse being a late transition state

for the forward reaction, CO binding to [(L)Cu^I]⁺), where

the solvent has no effect on the reaction (Scheme 4).

Summary of Kinetic Data and Activation Parameters.

Through transient absorbance laser flash photolysis, the rate

of CO association (k_CO) and dissociation (k_-CO) from copper-

(I) upon photodissociation has been quantified, along with

the activation parameters that govern this process. As the

electron-donating ability of the ligand increases such as in

1^R, the CO-binding rate (k_CO) increases and the CO dissocia-

tion rate (k_-CO) decreases. In a less-coordinating solvent, the

decreased enthalpic barrier (∆H^q) for k_COindicates that de-

ligation of CH₃CN preceeds CO rebinding. Activation

entropies (∆S^q) for k_COindicate a late transition state, as

Determination of k_-CO(s^-1) and Corresponding Activa-

tion Parameters. From the experimentally derived k_COand

K

_COvalues, the CO dissociation rate constant k_-CO(s^-1), and

its corresponding activation parameters can be calculated,

(eq 1, Experimental Section). Collected data support a simple

mechanism for CO de-ligation (i.e., the k_-COprocess) from

[(L)Cu^I(CO)]⁺complexes (Scheme 1). Within the R-TMPA

series 1^R, the rate constant for CO dissociation (k_-CO) 1.79

× 10⁴s^-1) is smallest for the most thermodynamically stable

complex, [Cu^I(NMe₂-tmpa)(CO)]⁺(1^Me2N-CO) in CH₃CN

(Table 3). As the ligand becomes less electron donating from

NMe₂- to MeO- to H-, the rate constant (k_-CO) becomes

progressively larger. Thus, breakage of the Cu-CO bond is

more difficult for the more-stable 1^Me2N-CO when compared

to that for [Cu^I(MeO-tmpa)(CO)]⁺(1^MeO-CO) and [Cu^I-

(tmpa)(CO)]⁺(1^H-CO). Further, the difficulty in breaking

the Cu-CO bond (k_-CO) is evident by the very large ∆H^q

terms for 1^MeO-CO (60.7 kJ mol^-1) and 1^H-CO (43.6 kJ

mol^-1) in CH₃CN.

supported by the activation parameters (∆H^q, ∆S^q) for k_-CO

.

The inability to dissociate CO from 2^Rsuggests that a strict

N₃-coodination mode also only possessing five-membered

chelate rings is necessary for photodissociation. The bulky

quinolyl arms of BQPA that induce the desirable N₃-

coodination mode in 3^bqpasupport this conclusion through a

lower k_COactivation enthalpy (∆H^q) in comparison to 1^H.

Inorganic Chemistry, Vol. 47, No. 1, 2008 253

Fry et al.

Scheme 5

Table 4. Comparison of O₂and CO Rate and Binding Constants at 298 K for [Cu^I(H-tmpa)(Solvent)]⁺(1^H), Hemocyanin (Hc), and Selected Hemes

compound

K

_CO(M^-1

)

K_O2(M^-1

)

k_CO(M^-1s^-1

)

k_O2(M^-1s^-1

)

k_-CO(s^-1

)

k_-O2(s^-1

)

[Cu^I(H-tmpa)(THF)]⁺(1^H-CO)

[Cu^I(H-tmpa)(CH₃CN)]⁺(1^H-CO)

hemocyanin (Hc)⁷

1.25 × 10⁵

15.4

0.38

1.9 × 10⁹

1.3 × 10⁹

1.54 × 10⁴1.3 × 10⁸

220

5.9 × 10⁷

5.8 × 10⁷

2.68 × 10⁵1.5 × 10⁸

(0.27 - 22) × 10⁴(0.18 - 90) × 10⁵

(2.0 - 7.7) × 10⁵(1.1 - 154) × 10⁶3-75

2.4 - 11.5

3.9 × 10³

1.9 × 10⁵

22

Fe(PF3CUIm) ^a,40

2.07 × 10⁹

1.45 × 10⁸

2.6 × 10⁷

6.7 × 10⁴

2.9 × 10⁷

4.8 × 10⁷

7.6 × 10⁵

2.6 × 10⁸

0.014

0.33

0.022

0.009

Fe(PF3CUPy) ^a,40

1.6 × 10³

3.0 × 10⁸

myoglobin - human (Mb) ³⁹

(0.74 - 117) × 10⁴

(1.4 - 25) × 10⁷

Hemoglobin - human, R state (Hb) ³⁹4.6 × 10⁸

(2.87 - 47.6) × 10⁵4.6 × 10⁶

(2.9 - 22) × 10⁷

13.1

^aFe(PF3CUIm) ) meso-5R,10R,15R-tris(o-pivalamidophenyl)-20â-{o-{[3-(N-imidazolyl)propyl] ureido}phenyl}porphyrinatoiron; Fe(PF3CUPy) ) meso-

5R,10R,15R-tris(o-pivalamidophenyl)- 20â-{o-{[3-(3-pyridyl)propyl]ureido}phenyl}porphyrinatoiron.

Photodissociation of Small Molecules from Copper

Proteins. As stated, the measurement of small molecule (CO,

O₂, or NO) recombination rates via photolytic methods is a

novel process for synthetic and biological copper(I) com-

plexes. However, work by Gorren et al. indicated that

photolysis of the nitrogen monoxide (NO) adduct of the

type-1 blue copper protein, azurin (Az), resulted in photo-

dissociation.⁷⁷Formation of AzNO led to the disappearance

of the g ) 2 EPR signal and the signature 625 nm absorbance

band. Upon photolysis, the original absorbance band reap-

pears but not the EPR signal, leading to the suggestion that

a photolabile Cu^I-NO⁺species was formed. Further inves-

tigation by Ehrenstein et al. involving temperature depen-

dence of this process found that binding of NO to azurin is

similar in nature to CO rebinding to myoglobin.⁷⁸At

temperatures less than 200 K, NO did not fully dissociate

from AzNO upon photolysis, resulting in geminate recom-

bination. Above 200 K, internal rebinding ceases as protein

fluctuations become fast, resulting in NO escaping the protein

matrix. In addition, such fluctuations decreased the internal

barrier for ligand rebinding, leading to exponential binding

of NO from the solvent. In contrast to Gorren et al.,

Ehrenstein and co-workers suggest that the photolabile

species was a tetragonal Cu^II-NO complex due to the growth

of a new absorption band below 400 nm upon binding NO.

time scale for Hc by Floyd et al.⁸⁰It is important to note

that the likely primary interaction of the protein dicopper(I)

site and dioxygen (i.e, formation of a Cu-O₂‚‚‚Cu^Ispecies)

was never observed, reiterating the significance of gaining

new insights of copper(I) interactions with small molecules

through kinetic and thermodynamic analyses of well-defined

coordination complexes.

Comparison of CO and O₂Binding to Copper Com-

plexes, Hemocyanin, and Hemes. Previously, we utilized

the ability to photodissociate CO from 1^H-CO to determine

the rate of O₂-binding to 1^H(Scheme 5).³⁸In the present

investigation, CO-binding to copper(I) complexes were

studied as a redox-inactive surrogate. We find that CO indeed

acts as an excellent O₂surrogate because the results with

regard to solvent interaction are nearly identical.^38,42For both

CO and O₂binding to 1^H, the enthalpy of activation (∆H^q)

is significantly lowered in THF when compared to that of

an organonitrile solvent. The off-rate of the small molecule

(Cu-CO or Cu-O₂bond cleavage) is unaffected by the

coordinating ability of the solvent, indicative of a dissociative

mechanism. Of greater interest is the nearly identical rate

constants obtained for the fast room-temperature binding of

O₂(k_O2(THF) ) 1.3 × 10⁹M^-1s^-1extrapolated from the

low-temperature data) and CO (k_CO(THF) ) 1.9 × 10⁹M^-1

s^-1, actual data obtained at room temperature) in the two

solvents (Table 4).^38,42However, the off-rate for O₂release

Photodissociation of O₂has been recently observed for

oxytyrosinase (Tyr_oxy)⁷⁹and oxyhemocyanin (Hc_oxy).⁸⁰Hirota

et al. report that 355 nm pulse irradiation of Tyr_oxyleads to

the bleaching of the 345 nm absorbance band on a micro-

second time scale, attributed to the dissociation of O₂from

the µ-η²:η²-peroxodicopper(II) center.⁷⁹Intensity recovery

corresponding to the rebinding of O₂to the deoxy enzyme

occurs on a millisecond time scale. Geminate recombination

of O₂was not observed for Tyr but was observed on a 80 ps

-

from [Cu^II(tmpa)(O₂)]⁺(k_-O2) 1.3 × 10⁸s^-1) in THF is

significantly greater than that for CO due to the large,

unfavorable entropy of activation term (∆S^q) 105 J mol^-1

K^-1for the O₂reaction,³⁸compared to -0.7 J mol^-1K^-1

for CO, Table 3) that most likely arises from the overall

change in geometry; loss of CO [Cu^I(tmpa)(CO)]⁺to give

[Cu^I(tmpa)(thf)]⁺involves re-ligation of the dangling pyridyl

arm (Scheme 2). Overall, CO-binding to copper(I) is highly

thermodynamically favored compared to O₂(Table 4) but

with respect to kinetics, the two small molecules appear to

be nearly identical. This explains why in our previous study

we observed highly competitive binding of CO and O₂to

1^H, with the overall thermodynamics favoring 1^H-CO; in

(77) Gorren, A. C. F.; Deboer, E.; Wever, R. Biochim. Biophys. Acta 1987,

916, 38-47.

(78) Ehrenstein, D.; Nienhaus, G. U. Proc. Natl. Acad. Sci. U.S.A. 1992,

89, 9681-9685.

(79) Hirota, S.; Kawahara, T.; Lonardi, E.; deWaal, E.; Funasaki, N.;

Canters, G. W. J. Am. Chem. Soc. 2005, 127, 17966-17967.

(80) Floyd, J. S.; Haralampus-Grynaviski, N.; Ye, T.; Zheng, B.; Simon,

J. D.; Edington, M. D. J. Phys. Chem. B 2001, 105, 1478-1483.

254 Inorganic Chemistry, Vol. 47, No. 1, 2008

Copper(I) Pyridylalkylamine Compounds

-

the presence of CO, [Cu^II(tmpa)(O₂)]⁺equilibrates back to

(1^H-CO) via MLCT (Cu^I(d₁₀)-N_py(π*)) excitation was

established.³⁸Now, the properties that govern this process

have been systematically investigated through the variation

of the ligand environment around the copper(I) metal center.

Characterization of the carbonyl adducts of 1-3 was

accomplished via X-ray crystallographic studies, cyclic

voltammetry, along with infrared (solid and solution) and

UV-vis spectroscopies. In addition, various experimental

techniques were utilized to determine equilibrium binding

constants (K_CO), association (k_CO) and dissociation (k_-CO) rate

constants, as well as the thermodynamic (∆H°, ∆S°) and

activation parameters (∆H^q, ∆S^q) that regulate this process.

First, a group of tetradentate ligand copper-CO species

(Chart 1) were structurally characterized. Through solid and

solution analyses, it was determined that the photolabile

carbonyl species in solution possesses a tridentate coordina-

tion mode to the copper ion with one dangling arm of the

tripodal ligand.

[Cu^I(tmpa)(CO)]⁺, allowing the determination of k_slow(Scheme

5).³⁸

Carbon monoxide binding constants (K_CO) for the dicop-

per(I) site in hemocyanin (Hc)⁷are comparable to that

reported here for mononuclear copper complexes, Table 4.

However, O₂-binding is significantly enhanced (>10⁴M^-1)

in the case of Hc and can easily be attributed to the dicopper

center of Hc stabilizing the dioxygen adduct, a peroxo

dicopper(II) complex where both copper ions are involved;

CO only binds to one of the copper centers.^9,81-85Thus,

comparisons of CO-binding to these copper complexes and

Hc are more relevant than O₂-binding. The rate at which

CO binds to copper is greatly enhanced in the case of our

copper complexes versus Hc (Table 4). These k_COvalues

are however affected by several factors including but not

limited to redox potentials and the ligand/protein environ-

ment. Yet, they also exhibit greater rates of CO dissociation

(k_-CO) than those exhibited by Hc. The higher dissociation

rates for both CO and O₂from these copper(I) complexes in

comparison to copper and heme proteins (Table 4 and

discussion above) are not only due to the more energy needed

to break the protein ligand-metal bond but also are due to

the greater complexity of the protein matrix.

Heme proteins and model systems have been studied

extensively and commonly employ carbon monoxide as a

redox-inactive surrogate to dioxygen.^{28-30,34,35,39-41,86,87}The

results presented here showing that copper rapidly reacts with

both CO and O₂are in line with studies carried out on

synthetic heme species. However, the copper(I) complexes

bind dioxygen and carbon monoxide at a much faster rate

than hemes but not with as high of an affinity. This

phenomenon is relevant to the mechanism of ligand exchange

in CcO, displaying that Cu_Bplays a regulatory role in the

kinetics of binding small molecules to the CcO heme/copper

active site.^28-30,86,87The first synthetic model system, [(⁶L)-

Fe^II(CO)‚‚‚Cu^I]⁺, directed at mimicking the CO transfer

chemistry of the heme_a3/Cu_Bactive site of CcO was reported

from our lab.⁸⁸This binucleating ligand system (⁶L) is

composed of a tetradentate chelate for copper (analogous to

1^H) covalently appended to the periphery of a difluorophenyl-

substituted porphyrinate. Like CcO, photodissociation of CO

from [(⁶L)Fe^II(CO)‚‚‚Cu^I]⁺results in the initial transfer of

CO to the copper(I) moiety forming [(⁶L)Fe^II‚‚‚Cu^I(CO)]⁺

before thermally equilibrating back to the initial species.⁸⁸

Summary/Conclusions. In previous work, the ability to

photodissociate carbon monoxide from [Cu^I(tmpa)(CO)]⁺

Second, it was elucidated how ligand electronic effects

and geometry dictate the strength in which CO binds to the

cuprous metal. When the electron-donating ability of the

tripodal, tetradentate ligand is increased, greater π back-

bonding results in stronger Cu-CO bonds, leading to K_CO

values on the order 1^Me2N-CO > 1^MeO-CO > 1^H-CO.

With systematic synthetic expansion of the five-membered

chelate rings like 1^Rto six-membered chelate rings like 2^R,

the stability of the CO adduct decreases, 1^H-CO > 2^pmea

-

CO > 2^pmap-CO. Because of the more-favorable tetrahedral

ligation of the PMEA and PMAP ligands to the copper alone,

2^Rare less reactive as reflected by their increased E_1/2values.

3^bqpadid not follow any trends because of its bulkier ligand

framework, which consistently favors an N₃coordination

mode. By contrast, 1^R-2^Rexhibit both N₃and N₄coordina-

tion modes in the absence of CO.

Third, CO photodissociated from 1^Me2N-CO, 1^MeO-CO,

1^H-CO, and 3^bqpa-CO in coordinating (CH₃CN) and

weakly coordinating (THF) solvent but not from 2^pmea-CO

and 2^pmap-CO. Each of the compounds that displayed CO

photodissociation behavior possess an N₃coordination mode

forming strictly five-membered chelate rings. In CH₃CN, no

particular trend in relation to the ligand environment was

observed when following the post-photolytic bimolecular

recombination of CO (k_CO) to the cuprous metal, where CO

+ 1^Me2N> 1^H> 1^MeO> 3^bqpa. In THF, bimolecular

recombination was observed where CO + 1^H> 3^bqpa

.

Finally, values for the dissociation of CO from copper-

(I)-CO complexes (k_-CO) k_CO/K_CO) were derived with a

clear trend with respect to electron-donating ability, where

k_-COfor 1^Me2N-CO < 1^MeO-CO < 1^H-CO. This simply

means that the terminal carbonyl complexes are more

favored, that is, the equilibrium lies to the left, as CO-release

becomes increasingly more difficult with greater ligand

electron-donating ability, resulting in stronger Cu-CO

complexes.

These findings offer clues as to how CO photodissociates

from [(L)Cu^I(CO)]⁺complexes and how copper binds small

molecules (CO and O₂) in relation to other biologically

relevant copper complexes as well as proteins. Such insights

(81) Root, R. W. J. Biol. Chem. 1934, 104, 239-244.

(82) Peisach, J.; Aisen, P.; Blumberg, W. E. The Biochemistry of Copper;

Academic Press: New York, 1966.

(83) Fager, L. Y.; Alben, J. O. Biochemistry 1972, 11, 4786.

(84) Connelly, P. R.; Gill, S. J.; Miller, K. I.; Zhou, G.; Vanholde, K. E.

Biochemistry 1989, 28, 1835-1843.

(85) Connelly, P. R.; Johnson, C. R.; Robert, C. H.; Bak, H. J.; Gill, S. J.

J. Mol. Biol. 1989, 207, 829-832.

(86) Fiamingo, F. G.; Altschuld, R. A.; Moh, P. P.; Alben, J. O. J. Biol.

Chem. 1982, 257, 1639-1650.

(87) Dyer, R. B.; Einarsdottir, O.; Killough, P. M.; Lopezgarriga, J. J.;

Woodruff, W. H. J. Am. Chem. Soc. 1989, 111, 7657-7659.

(88) Fry, H. C.; Cohen, A. D.; Toscano, J. P.; Meyer, G. J.; Karlin, K. D.

J. Am. Chem. Soc. 2005, 127, 6225-6230.

Inorganic Chemistry, Vol. 47, No. 1, 2008 255

Fry et al.

include the possible transient binding of small molecules to

the Cu_Bactive site of CcO prior to binding to the heme_a3

active site. Carbon monoxide binds faster to these copper(I)

complexes than most hemes, but thermodynamics favor

heme-CO species.

Health, Grant GM28962; G.J.M., National Science Founda-

tion, Grant CHE 0616500).

Supporting Information Available: Additional cyclic volta-

mmetry, infrared spectra, ¹H NMR spectra, spectrophotometric

titrations, Van’t Hoff plots, transient absorption difference spectra,

linear [CO] dependence, Eyring plots, crystal structures and CIF

files for crystal structures. This material is available free of charge

via the Internet at http://pubs.acs.org.

Acknowledgment. We are very grateful for financial

support of this research (K.D.K., National Institutes of

IC701903H

256 Inorganic Chemistry, Vol. 47, No. 1, 2008

Article Doi

Carbon monoxide coordination and reversible photodissociation in copper(I) pyridylalkylamine compounds

DOI: 10.1021/ic701903h

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ic701903h

Full text of DOI:10.1021/ic701903h

Products guided by the article

R&D Labs maybe for 1004985-31-2

Relevant to this article

Hot Product