New Monodentate Amidine Superbasic Ligands with a Single Configuration in fac-[Re(CO)₃(5,5′- or 6,6′-Me₂bipyridine)(amidine)]BF₄ Complexes

Abstract

Treatment of two precursors, fac-[Re(CO)₃(L)(CH₃CN)]BF₄ [L = 5,5′-dimethyl-2,2′-bipyridine (5,5′-Me₂bipy) (1) and 6,6′-dimethyl-2,2′-bipyridine (6,6′-Me₂bipy) (2)], with five C₂-symmetrical saturated heterocyclic amines yielded ten new amidine complexes, fac-[Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ [Y = CH₂, (CH₂)₂, (CH₂)₃, NH or O]. All ten complexes possess the novel feature of having only one isomer (amidine E configuration), as established by crystallographic and ¹H NMR spectroscopic methods. We are confident that NMR signals of the other possible isomer (amidine Z configuration) would have been detected, if it were present. Isomers are readily detected in closely related amidine complexes because the double-bond character of the amidine C–N3 bond (N3 is bound to Re) leads to slow E to Z isomer interchange. The new fac-[Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes have C–N3 bonds with essentially identical double-bond character. However, the reason that the Z isomer is so unstable as to be undetectable in the new complexes is undoubtedly because of unfavorable clashes between the equatorial ligands and the bulky N(CH₂CH₂)₂Y ring moiety of the axial amidine ligand. The amidine formation reactions in acetonitrile (25 °C) proceeded more easily with 2 than with 1, indicating that the distortion in 6,6′-Me₂bipy resulting from the proximity of the methyl substituents to the inner coordination sphere enhanced the reactivity of the coordinated CH₃CN. Reaction times for 1 and 2 exhibited a similar dependence on the basicity and ring size of the heterocyclic amine reactants. Moreover, when the product of the reaction of 1 with piperidine, fac-[Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄, was challenged in acetonitrile-d₃ or CDCl₃ with a fivefold excess of the strong 4-dimethylaminopyridine ligand, there was no evidence for replacement of the amidine ligand after two months, thus establishing that the piperidinylamidine ligand is a robust ligand. This chemistry offers promise as a suitable means for preparing isomerically pure conjugated fac-[^99mTc(CO)₃L]^n+/− imaging agents, including conjugates with known bioactive heterocyclic amines.

Introduction

Owing to the many ideal properties of the fac-[M^I(CO)₃] core in radiopharmaceuticals, fac-[M^I (CO)₃L]ⁿ (M = various isotopes of Tc and Re)complexes have recently been receiving much attention.^1–7 Some fac-[^99mTc^I(CO)₃L]ⁿ imaging agents have undergone evaluation in humans,^8–9 and fac-[^186/188Re^I(CO)₃L]ⁿ agents are emerging as being among the most promising radionuclides for therapeutic applications.^2,10–11 At present, great interest surrounds the concept of combining ^99mTc and ^186/188Re with biomolecules in order to produce selective targeting agents.^{5–6,11–17} fac-[Re^I(CO)₃L]ⁿ complexes prepared with natural-abundance rhenium are excellent models for the short-lived fac-[M^I(CO)₃L]ⁿ radiopharmaceuticals and are almost non-radioactive. Thus, the investigation of fac-[Re^I(CO)₃L]ⁿ complexes both aids in interpreting the chemistry of the radiopharmaceuticals and offers the potential for the discovery of new chemistry, some of which could be applied to radiopharmaceutical development.^18–19

Our objectives are aimed at expanding the known chemistry of complexes with the fac-[Re^I(CO)₃] core.^7,20–21 Syntheses in aqueous media carried out with the commonly used precursor, aqueous fac-[Re^I(CO)₃(H₂O)₃]^+,22 have some limitations.²³ Thus, we have recently investigated the suitability of fac-[Re(CO)₃(CH₃CN)₃]X (X = PF₆ or BF₄) as a precursor for the synthesis of new complexes in organic solvents.²³ Treatment of fac-[Re(CO)₃(CH₃CN)₃]X with bidentate aromatic sp² N-donor bipyridine-type L in either acetonitrile or benzene as solvent produced the desired fac-[Re(CO)₃(L)(CH₃CN)]X complexes in excellent yield [e.g., when L = 2,2′-bipyridine (bipy) or a dimethyl-2,2′-bipyridine (Me₂bipy), Scheme 1].²⁴ However, a recent study revealed that reactions to form these complexes in methanol instead led to addition of solvent to bound acetonitrile, forming iminoether complexes, fac-[Re(CO)₃(Me₂bipy)(HNC(CH₃)OCH₃)]BF₄.²³ The original acetonitrile carbon with a triple bond to the rhenium-bound nitrogen (N3) is converted in the reaction to an iminoether carbon (C_ie), and N3 adds a proton and rehybridizes from sp to sp² (Scheme 1). The C_ie–N3 bond has double-bond character, and the iminoether ligand potentially can have E and Z configurations. However, the Z isomer (Scheme 1) is favored exclusively because the axial iminoether ligand steric repulsions with the equatorial ligands (the two CO’s and the Me₂bipy) are lower for the Z configuration than for the E configuration.²³

Scheme 1.

General reaction scheme for the synthesis of [Re(CO)₃(L)(CH₃CN)]⁺ starting material²⁴ and for the formation of [Re(CO)₃(Me₂bipy)(HNC(CH₃)OCH₃)]⁺ (iminoether) complexes.²³

The reactions of fac-[Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]⁺ (1) with alcohols to form iminoethers were slow.²³ On the other hand, the related reactions of primary amines with 1 to form amidine complexes, fac-[Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NHR)]⁺, were more rapid.²⁵ However, these amidine complexes exist as mixtures of isomers. In the HNC(CH₃)NHR ligands, both C–N bonds involving the amidine carbon (C_am), C_am–N3 and C_am–N4, have double-bond character. This situation raises the possibility that four configurations (E, E′, Z and Z′) of the amidine ligands could exist (Figure 1). In fact, three configurations (E, E′, and Z) were found.²⁵ The isomers are named using these configurations. As illustrated and discussed below, steric effects strongly influence the relative abundance of the isomers.

Figure 1.

The four conceivable [Re(CO)₃(5,5′-Me₂bipy)HNC(CH₃)NHR)]⁺ isomers, in which N–N denotes the 5,5′-Me₂bipy ligand. The isomers with the E′ and Z configurations are typically abundant. The isomer with the Z′ configuration is unstable and not observed.²⁵ The isomer with the E configuration is known, but its abundance is usually too low to allow detection. However, as illustrated here, the pathway between the E′ and Z configurations undoubtedly passes through the E configuration and not the Z′ configuration.

The amidine group, such as that present in fac-[Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NHR)]BF₄ complexes,²⁵ has the potential to serve as a linking group in the conjugation of the fac-[M(CO)₃]⁺ core (M = ^99mTc and ^186/188Re radionuclides) with biomedical targeting moieties. The nitrogen donor group in amidine (and iminoether) ligands is superbasic.^23,25 However, the finding of isomers of these complexes (Figure 1) complicates the development of agents useful for biomedical imaging. Therefore, we now explore amidine ligands with a C₂-symmetrical NR₂ substituent in place of the NHR substituent. This change eliminates the possibility of two configurations about the C_am–N4 bond, restricting the number of conceivable isomers to two (with E or Z configurations about the C_am–N3 bond). Furthermore, we expected that a large difference in substituent bulk (NR₂ vs. CH₃) on C_am should favor the E isomer exclusively.

We chose C₂-symmetrical saturated heterocyclic secondary amines in our synthetic strategy because many related symmetric heterocyclic amine derivatives are present in ^99mTc and ^186/188Re agents^13,26–31 and in successful drugs.^31–32 Because their use as ubiquitous building blocks in the synthesis of pharmaceuticals³¹ has provided information on the synthesis and properties of such amines, these amines are particularly desirable candidates for study. Indeed, a modified arylpiperazine was employed in one of the earliest examples of a fac-[^99mTc(CO)₃]⁺-containing agent linked to a targeting biomolecule.¹³ All of the new complexes discussed below have the facial geometry, and thus from this point onward we omit the fac- designation when discussing specific compounds.

Experimental Section

Starting Materials

Re(CO)₅Br was synthesized as described in the literature.³³ Re₂(CO)₁₀, 5,5′-dimethyl-2,2′-bipyridine (5,5′-Me₂bipy), 6,6′-dimethyl-2,2′-bipyridine (6,6′-Me₂bipy), piperidine, homopiperidine, heptamethyleneimine, morpholine, piperazine, and AgBF₄ were obtained from Aldrich. [Re(CO)₃(CH₃CN)₃]BF₄ (prepared by a slight modification of a known procedure³⁴) was used to prepare [Re(CO)₃(5,5′- or 6,6′-Me₂bipy)(CH₃CN)]BF₄.²⁴

NMR Measurements

¹H NMR spectra were recorded on a 400 MHz Bruker spectrometer. Peak positions are relative to TMS or to solvent residual peak, with TMS as reference. All NMR data were processed with TopSpin and Mestre-C software.

X-ray Data Collection and Structure Determination

Intensity data were collected at low temperature on a Nonius Kappa CCD diffractometer fitted with an Oxford Cryostream cooler with graphite-monochromated Mo Kα (λ = 0.71073 Å) radiation. Data reduction included absorption corrections by the multi-scan method, with HKL SCALEPACK.³⁵ All X-ray structures were determined by direct methods and difference Fourier techniques and refined by full-matrix least squares by using SHELXL-97.³⁶ All non-hydrogen atoms were refined anisotropically. All H atoms were visible in difference maps, but were placed in idealized positions, except for N-H hydrogen atoms, for which coordinates were refined. A torsional parameter was refined for each methyl group. For compounds 4, 5, and 9, the BF₄⁻ site was shared by a few percent bromide, and the occupancies of the two anions were constrained to sum to unity in the refinement. In compound 10, the BF₄⁻ is disordered into two orientations and the 8-membered ring is disordered into two conformations. The occupancies refined to 0.891(5):0.109(5) for the anion and 0.521(6):0.479(6) for the 8-membered ring. Crystal data and details of refinements are listed in Tables 1 and 2.

Table 1.

Crystal Data and Structural Refinement for Complexes Having the General Formula, [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄

Y = Complex	CH2 3	(CH2)2 4	(CH2)3 5	NH 6	O 7
empirical formula	C22H26N4O3Re·BF4	C23H28N4O3Re·0.95(BF4)·0.05(Br)	C24H30N4O3Re·0.96(BF4)·0.04(Br)	C21H25N5O3Re·BF4	C21H24N4O4Re·BF4
Fw	667.48	681.16	695.50	668.47	669.45
crystal system	monoclinic	monoclinic	monoclinic	monoclinic	monoclinic
space group	P21/n	P21/n	P21/n	P21/n	P21/n
a (Å)	11.4576(10)	13.3550(15)	13.7247 (14)	11.6155(10)	11.3847(9)
b (Å)	13.4757(15)	13.2081(14)	11.1284 (10)	12.9640(14)	13.3112(10)
c (Å)	15.9875(15)	14.7562(18)	18.141 (2)	15.8176(11)	15.7988(15)
β (deg)	97.502(5)	105.347(6)	109.392 (3)	97.341(6)	97.843(6)
V (Å3)	2447.3(4)	2510.1(5)	2613.6 (5)	2362.3(4)	2371.8(3)
T (K)	200	150	95	95	90
Z	4	4	4	4	4
ρcalc (Mg/m3)	1.812	1.802	1.768	1.880	1.875
abs coeff (mm−1)	5.03	4.98	4.78	5.21	5.19
2θmax (°)	60.2	61.0	72.6	68.4	70.0
R [I> 2σ(I)]a	0.032	0.032	0.029	0.032	0.033
wR2b	0.073	0.075	0.060	0.067	0.075
w scheme d,e	0.0315, 2.4602	0.0343, 3.5449	0.0230, 1.6351	0.0252, 1.6969	0.0354, 0
data/param	7180/323	7045/344	12259/345	9257/326	10062/323
Res. dens (eÅ−3)	1.23, −1.23	1.08, −1.85	1.16, −1.54	1.37, −1.72	1.46, −1.70

^aR = (Σ||F_o| − |F_c||)/Σ|F_o|.

^bwR2 = [Σ[w(F_o² − F_c²)²]/Σ[w(F_o²)²]]^1/2, in which w = 1/[σ²(F_o²) + (dP)² + (eP)] and P = (F_o² + 2F_c²)/3.

Table 2.

Crystal Data and Structural Refinement for Complexes Having the General Formula, [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄

Y = Complex	CH2 8	(CH2)2 9	(CH2)3 10	NH 11	O 12
empirical formula	C22H26N4O3Re·BF4	C23H28N4O3Re·0.97(BF4)·0.03(Br)	C24H30N4O3Re·BF4	C21H25N5O3Re·BF4	C21H24N4O4Re·BF4
Fw	667.48	681.31	695.53	668.47	669.45
crystal system	monoclinic	monoclinic	monoclinic	monoclinic	monoclinic
space group	P21/n	P21/n	P21/c	P21/n	P21/n
a (Å)	8.9242(5)	11.4040(10)	12.5348(10)	12.401(2)	15.6441(15)
b (Å)	21.862(2)	17.724(2)	10.8431(9)	14.221(3)	9.3838(10)
c (Å)	12.3282(10)	12.3056(11)	19.692(2)	13.630(2)	16.1555(12)
β (deg)	95.358(4)	98.914(5)	105.942(4)	100.750(9)	91.294(4)
V (Å3)	2394.7(3)	2457.2(4)	2573.5(4)	2361.5(7)	2371.0(4)
T (K)	90	100	90	95	90
Z	4	4	4	4	4
ρcalc (Mg/m3)	1.851	1.842	1.795	1.880	1.875
abs coeff (mm−1)	5.14	5.05	4.79	5.21	5.19
2θmax (°)	72.0	69.8	68.0	65.2	71.4
R [I> 2σ(I)]a	0.031	0.031	0.030	0.037	0.029
wR2b	0.074	0.068	0.070	0.097	0.069
w scheme d,e	0.0349, 0	0.0271, 2.2079	0.0307, 2.8985	0.0575,0.7244	0.0349, 1.1307
Data/param	10749/323	10050/333	9861/388	8537/326	10909/323
Res. dens (eÅ−3)	1.59, −2.07	1.35, −1.99	1.83, −1.58	4.45, −2.55	2.22, −1.69

^aR = (Σ||F_o| − |F_c||)/Σ|F_o|.

^bwR2 = [Σ[w(F_o² − F_c²)²]/Σ[w(F_o²)²]]^1/2, in which w = 1/[σ²(F_o²) + (dP)² + (eP)] and P = (F_o² + 2F_c²)/3.

General Synthesis of Amidine Complexes

An acetonitrile solution (6 mL) of [Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]BF₄ (1) or [Re(CO)₃(6,6′-Me₂bipy)(CH₃CN)]BF₄ (2) (40 mg, 0.06 mmol) was treated with an amine (0.60 mmol), and the reaction mixture was stirred at room temperature for 30 min or as specified. The volume was reduced to ~1 mL by rotary evaporation. Addition of diethyl ether to the point of cloudiness (~10–200 mL) produced a yellow crystalline material that was collected on a filter, washed with diethyl ether, and dried. All ¹H NMR spectra recorded both immediately upon dissolution of the products (3 to 12) and subsequently showed signals for one isomer.

The ¹H NMR spectrum of all crystals described below was identical to that of the product obtained by this procedure. In order to study the progress of the amidine formation reactions, a 10 mM solution of 1 or 2 was prepared in 600 μL of acetonitrile-d₃. We refer to such a solution as the 10 mM solution. An excess of amine (100 mM) was added to the 10 mM solution, and the reaction was monitored by NMR spectroscopy. In all cases, the only signals observed for products were those expected from the isolated products.

Synthesis of [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ (3)

The use of this general method in the reaction of 1 with piperidine (59 μL, 0.60 mmol) afforded 30 mg (74% yield) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.85 (s, 2H, H6/6′), 8.26 (d, J = 8.4 Hz, 2H, H3/3′), 8.04 (d, J = 8.4 Hz, 2H, H4/4′), 4.78 (b, 1H, NH), 3.01 (m, 4H, 2CH₂), 2.48 (s, 6H, 5/5′-2CH₃), 2.10 (s, 3H, CCH₃), 1.48 (m, 2H, CH₂), 1.28 (m, 4H, 2CH₂).

X-ray quality crystals of 3 (E isomer) were produced upon slow evaporation of a solution of the crystalline material (5 mg/6 mL) in a 1/5 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 1 with piperidine (5.9 μL) as described above indicated that no signals for 1 remained after 5 min, and signals for 3 were the only product signals present.

Synthesis of [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₂)]BF₄ (4)

The use of the general method in the reaction of 1 with homopiperidine (60 μL, 0.60 mmol) produced 33 mg (80% yield) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.87 (s, 2H, H6/6′), 8.27 (d, J = 8.6 Hz, 2H, H3/3′), 8.05 (d, J = 8.1 Hz, 2H, H4/4′), 4.52 (b, 1H, NH), 3.30 (b, m, 2H, CH₂), 2.96 (b, m, 2H, CH₂), 2.48 (s, 6H, 5/5′-2CH₃), 2.10 (s, 3H, CCH₃), 1.46 (b, m, 2H, CH₂), 1.31 (b, m, 2H, CH₂), 1.15 (b, m, 2H, CH₂), 0.96 (b, m, 2H, CH₂).

X-ray quality crystals of 4 (E isomer) grew upon slow evaporation of a solution of the crystalline material (5 mg/4 mL) in a 1/3 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 1 with homopiperidine (6 μL) as described above indicated that no signals for 1 remained after ~8 min.

Synthesis of [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₃)]BF₄ (5)

The use of the general method in the reaction of 1 with heptamethyleneimine (76 μL, 0.60 mmol), but stirring for 8 h, yielded 13 mg (32%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.87 (s, 2H, H6/6′), 8.27 (d, J = 8.3 Hz, 2H, H3/3′), 8.04 (d, J = 8.0 Hz, 2H, H4/4′), 4.49 (b 1H, NH), 3.25 (b, 2H, CH₂), 3.05 (b, 2H, CH₂), 2.48 (s, 6H, 5/5′-2CH₃), 2.12 (s, 3H, CCH₃), 1.50 (b, m, 2H, CH₂), 1.39(b, m, 2H, CH₂), 1.18 (b, m, 2H, CH₂), 0.96 (b, m, 2H, CH₂), 0.71 (b, m, 2H, CH₂). X-ray quality crystals of 5 (E isomer) grew upon slow evaporation of a solution of the crystalline material (10 mg/~200 mL) in a 1/200 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 1 with heptamethyleneimine (7.6 μL) as described above indicated that no signals for 1 remained after 6 h.

Synthesis of [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂NH)]BF₄ (6)

The general synthetic reaction of 1 with piperazine (52 mg, 0.60 mmol) yielded 34 mg (84%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.85 (s, 2H, H6/6′), 8.26 (d, J = 8.4 Hz, 2H, H3/3′), 8.04 (d, J = 8.4 Hz, 2H, H4/4′), 4.84 (b, 1H, NH), 2.95 (m, 4H, 2CH₂), 2.53 (m, 4H, 2CH₂), 2.48 (s, 6H, 5/5′-2CH₃), 2.12 (s, 3H, CCH₃).

X-ray quality crystals of 6 (E isomer) formed upon slow evaporation of a 16 mL solution of the crystalline material (5 mg) in a 1/15 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 1 with piperazine (5.2 mg) as described above indicated that no signals for 1 remained after 20 min.

Synthesis of [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂O)]BF₄ (7)

The general synthetic reaction of 1 with morpholine (53 μL, 0.60 mmol) (stirring time, 6 h) yielded 33 mg (83%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.85 (s, 2H, H6/6′), 8.26 (d, J = 8.4 Hz, 2H, H3/3′), 8.05 (d, J = 8.1 Hz, 2H, H4/4′), 4.94 (b, 1H, NH), 3.45 (m, 2H, CH₂), 3.00 (m, 2H, CH₂), 2.48 (s, 6H, 5/5′-2CH₃), 2.14 (s, 3H, CCH₃).

X-ray quality crystals of 7 (E isomer) grew upon slow evaporation of a 4 mL solution of the crystalline material (5 mg) in a 1/3 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 1 with morpholine (5.3 μL) as described above indicated that no signals for 1 remained after 4 h.

Synthesis of [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ (8)

The general treatment of 2 with piperidine (59 μL, 0.60 mmol) yielded 35 mg (88%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.19 (d, J = 8.1 Hz, 2H, H3/3′), 8.06 (t, J = 7.9 Hz 2H, H4/4′), 7.62 (d, J = 7.9 Hz, 2H, H5/5′), 5.14 (b, 1H, NH), 3.06 (s, 6H, 6/6′-2CH₃), 3.03 (overlapped m, 4H, 2CH₂), 1.60 (s, 3H, CCH₃), 1.53 (m, 2H, CH₂), 1.29 (m, 4H, 2CH₂).

X-ray quality crystals of 8 (E isomer) formed upon slow evaporation of a 6 mL solution of the crystalline material (5 mg) in a 1/5 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 2 with piperidine (5.9 μL) as described above indicated that no signals for 2 remained after 3 min.

Synthesis of [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₂)]BF₄ (9)

The general treatment of 2 with homopiperidine (60 μL, 0.60 mmol) yielded 32 mg (78%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.19 (d, J = 7.8 Hz, 2H, H3/3′), 8.06 (t, J = 7.9 Hz, 2H, H4/4′), 7.61 (d, J = 7.8 Hz, 2H, H5/5′), 4.90 (b, 1H, NH), 3.26 (b, m, 2H, CH₂), 3.07 (s, 6H, 6/6′-2CH₃), 3.04 (overlapped m, 2H, CH₂), 1.62 (s, 3H, CCH₃), 1.44 (b, m, 2H, CH₂), 1.38 (b, m, 2H, CH₂), 1.32 (b, m, 2H, CH₂), 1.11 (b, m, 2H, CH₂).

X-ray quality crystals of 9 (E isomer) grew upon slow evaporation of a 5 mL solution of the crystalline material (5 mg) in a 1/4 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 2 with homopiperidine (6 μL) as described above indicated that no signals for 2 remained after ~4.5 min.

Synthesis of [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₃)]BF₄ (10)

The general treatment of 2 with heptamethyleneimine (76 μL, 0.60 mmol) afforded 15 mg (35%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.19 (d, J = 8.1 Hz, 2H, H3/3′), 8.07 (t, J = 7.9 Hz, 2H, H4/4′), 7.63 (d, J = 7.7 Hz, 2H, H5/5′), 4.82 (b, 1H, NH), 3.23 (b, m, 2H, CH₂), 3.16 (b, m, 2H, CH₂), 3.07 (s, 6H, 6/6′-2CH₃), 1.66 (s, 3H, CCH₃), 1.49 (b, m, 2H, CH₂), 1.42 (b, m, 2H, CH₂), 1.32 (b, m, 2H, CH₂), 1.21 (b, m, 2H, CH₂), 0.92 (b, m, 2H, CH₂).

X-ray quality crystals of 10 (E isomer) formed upon slow evaporation of a solution of the crystalline material (5 mg/16 mL) in a 1/15 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 2 with heptamethyleneimine (7.6 μL) as described above indicated that no signals for 2 remained after 6 min.

Synthesis of [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂NH)]BF₄ (11)

The general treatment of 2 with piperazine (52 mg, 0.60 mmol) yielded 33 mg (83%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.19 (d, J = 8.0 Hz, 2H, H3/3′), 8.07 (t, J = 7.9 Hz, 2H, H4/4′), 7.62 (d, J = 7.7 Hz, 2H, H5/5′), 5.18 (b, 1H, NH), 3.05 (s, 6H, 6/6′-2CH₃), 2.96 (m, 4H, 2CH₂), 2.53 (m, 4H, 2CH₂), 1.63 (s, 3H, CCH₃).

X-ray quality crystals of 11 (E isomer) grew upon slow evaporation of a solution of the crystalline material (5 mg/5 mL) in a 1/4 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 2 with piperazine (5.1 mg) as described above indicated that no signals for 2 remained after 3 min.

Synthesis of [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂O)]BF₄ (12)

The general treatment of 2 with morpholine (53 μL, 0.60 mmol) (stirring time, 1 h) afforded 36 mg (90%) of yellow crystalline material. ¹H NMR signals (ppm) in acetonitrile-d₃: 8.19 (d, J = 7.9 Hz, 2H, H3/3′), 8.07 (t, J = 7.9 Hz, 2H, H4/4′), 7.62 (d, J = 7.7 Hz, 2H, H5/5′), 5.30 (b, 1H, NH), 3.45 (m, 4H, 2CH₂), 3.05 (s, 6H, 6/6′-2CH₃), 3.01 (m, 4H, 2CH₂), 1.66 (s, 3H, CCH₃).

X-ray quality crystals of 12 (E isomer) grew upon slow evaporation of a solution of the crystalline material (5 mg/4 mL) in a 1/3 (v/v) mixture of acetonitrile/diethyl ether. The ¹H NMR spectrum of the crystals dissolved in acetonitrile-d₃ was identical to that of the bulk product.

Monitoring the progress of the reaction of 2 with morpholine (5.3 μL) as described above indicated that no signals for 2 remained after 30 min.

Challenge Reactions

A 5 mM solution of [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ (3) in acetonitrile-d₃ (600 μL) was treated with a fivefold excess of 4-dimethylaminopyridine (2.0 mg, 25 mM), and the solution was monitored over time by ¹H NMR spectroscopy. A similar experiment was conducted in CDCl₃.

Results and Discussion

Synthesis

Treatment of [Re(CO)₃(L)(CH₃CN)]BF₄ (L = 5,5′-Me₂bipy (1), and 6,6′-Me₂bipy (2)) with heterocyclic amines in acetonitrile at room temperature afforded good yields (usually greater than 70%) of amidine complexes of the general formula, [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ (L = 5,5′-Me₂bipy or 6,6′-Me₂bipy; Y = CH₂, (CH₂)₂, (CH₂)₃, NH, or O), as illustrated in Figure 2. ¹H NMR spectroscopic studies and structural characterization by single-crystal X-ray crystallography (see below) show that the reactions with cyclic amines form only one isomer (E) of the new amidine complexes. Reactions are often rapid at ambient temperature (≤3 min for complete reaction). Because the greater reactivity of [Re(CO)₃(6,6′-Me₂bipy)(CH₃CN)]BF₄ (2) than of [Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]BF₄ (1) with a given amine is best understood after a discussion of structural and spectroscopic results, we shall return to the topic of reaction times later.

Figure 2.

Reactions forming [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]⁺ complexes observed upon treatment of [Re(CO)₃(L)(CH₃CN)]⁺ complexes with heterocyclic amines (HN(CH₂CH₂)₂Y) in acetonitrile at 25 °C.

Structural Results

Summarized in Tables 1 and 2 are the crystal data and details of the structural refinement for complexes 3-12, having the general formula, [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ (L = 5,5′-Me₂bipy or 6,6′-Me₂bipy, Y = CH₂, (CH₂)₂, (CH₂)₃, NH, or O). Figures 3 and 4 show the ORTEP plots of the cations in complexes 3-12, together with the numbering scheme used to describe the solid-state data. All complexes have a pseudo octahedral structure, in which the three carbonyl ligands are coordinated facially. The remaining three coordination sites are occupied by the two nitrogen atoms of L and by one nitrogen atom of the neutral monodentate amidine ligand having the E configuration. Ni(II) amidine complexes formed upon addition of secondary amines to coordinated acetonitrile have the E configuration in the solid state.^38–39

Figure 3.

ORTEP plots of the cations in [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ (3), [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₂)]BF₄ (4), [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃) N(CH₂CH₂)₂(CH₂)₃)]BF₄ (5), [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂NH)]BF₄ (6), and [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂O)]BF₄ (7). Thermal ellipsoids are drawn with 50% probability.

Figure 4.

ORTEP plots of the cations in [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ (8), [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₂)]BF₄ (9), [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₃)]BF₄ (10), [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂NH)]BF₄ (11), and [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂O)]BF₄ (12). Thermal ellipsoids are drawn with 50% probability. For 10, both conformations of the disordered 8-membered ring are shown, and H atoms are not illustrated, except for N-H.

The Re–C bond distances (not shown) of the two CO groups cis to the amidine ligand are generally not significantly different from the one trans to it in all complexes (3-12). All of the [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (3-7) show Re–N bond lengths (Table 3) comparable to the typical Re sp² nitrogen bond length, typically ranging from 2.14 to 2.18 Å.²² This result is consistent with the structural results for the recent monodentate amidine complexes of Re^I with primary amines.²⁵ As found for the iminoether complexes, in which the Re–N3 bond lengths found for [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)OCH₃)]BF₄ (2.1860(18) Å) and [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)OCH₃)]BF₄ (2.175(3) Å) were not significantly different,23 the Re-N3 bond lengths are quite similar for complexes 3-12. These bond lengths appear to be very slightly longer for the 6,6′-Me₂bipy complexes (range 2.1848(18) – 2.193(2), mean 2.190 Å) than for the 5,5′-Me₂bipy complexes (range 2.178(3) – 2.1806(18), mean 2.179 Å).

Table 3.

Selected Bond Distances (Å) and Angles (deg) for Complexes Having the General Formula, [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄

Y = complex	CH2 3	(CH2)2 4	(CH2)3 5	NH 6	O 7
bond distances
Re–N1	2.168 (3)	2.172 (3)	2.1691 (18)	2.177 (2)	2.177 (2)
Re–N2	2.186 (3)	2.173 (3)	2.1823 (18)	2.190 (2)	2.194 (2)
Re–N3	2.178 (3)	2.179 (3)	2.1806 (18)	2.179 (2)	2.178 (2)
N3–C16	1.306 (4)	1.310 (5)	1.308 (3)	1.300 (4)	1.304 (3)
N4–C16	1.346 (5)	1.344 (5)	1.346 (3)	1.354 (4)	1.359 (3)
bond angles
N1–Re–N2	75.06 (11)	74.58 (11)	74.82 (6)	75.08 (8)	75.16 (8)
N1–Re–N3	80.23 (11)	83.44 (11)	87.21 (6)	79.90 (9)	78.78 (8)
N2–Re–N3	86.41 (11)	79.02 (11)	79.34 (6)	86.11 (9)	86.30 (8)
Re–N3–H3N	113 (3)	111 (3)	106 (2)	110 (2)	110 (2)
Re–N3–C16	137.4 (2)	135.6 (3)	136.82 (15)	136.5 (2)	137.03 (18)
C16–N3–H3N	110 (3)	110 (3)	116 (2)	114 (2)	113 (2)
N3–C16–N4	123.6 (3)	122.9 (3)	122.47 (19)	123.3 (3)	122.5 (2)
N3–C16–C17	118.6 (3)	118.5 (3)	119.58 (19)	119.2 (3)	119.9 (2)
N4–C16–C17	117.8 (3)	118.7 (3)	117.95 (19)	117.4 (2)	117.7 (2)
C16–N4–C18	122.2 (3)	121.2 (3)	123.86 (18)	121.8 (2)	121.4 (2)
C16–N4–C(n)a	120.3 (3)b	122.7 (3)c	120.44 (18)d	120.7 (2)e	120.1 (2)f

^an varies in number according to the R group.

^bn = 22.

^cn = 23.

^dn = 24.

^en = 21.

^fn = 21.

The recent study of fac-[Re(CO)₃(L)(HNC(CH₃)OCH₃)]BF₄ complexes revealed that the Re–N bond lengths in the equatorial plane were significantly longer for L = 6,6′-Me₂bipy than for L = 5,5′-Me₂bipy.²³ These examples of a slight Re–N bond lengthening were attributed to the distorted nature of the 6,6′-Me₂bipy ligand as a result of the close proximity of the two methyl substituents to the equatorial carbonyl groups. A comparison of the equatorial Re–N bond lengths (Tables 3 and 4) of all five [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (8-12) with those of the corresponding five [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (3-7) reveals that only some bonds in the 6,6′-Me₂bipy complexes are slightly longer by criteria of statistical significance. However, as for the Re–N3 axial distances, the equatorial Re–N(Me₂bipy) distances appear to be on average very slightly longer for the 6,6′-Me₂bipy complexes (range 2.194(2) – 2.213(2), mean 2.206 Å) than for the 5,5′-Me₂bipy complexes (range 2.168(3) – 2.194(2), mean 2.179 Å). Thus, the more extensive solid-state results for complexes 3-12 now available indicate that the 6,6′-methyl groups in 8-12 affect the equatorial Re–N bond distances only slightly.

Table 4.

Selected Bond Distances (Å) and Angles (deg) for Complexes Having the General Formula, [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄

Y = complex	CH2 8	(CH2)2 9	(CH2)3 10	NH 11	O 12
bond distances
Re–N1	2.213 (2)	2.203 (2)	2.211 (2)	2.212 (3)	2.2051 (19)
Re–N2	2.1984 (18)	2.194 (2)	2.211 (2)	2.202 (3)	2.2086 (19)
Re–N3	2.193 (2)	2.188 (2)	2.190 (2)	2.192 (3)	2.1848 (18)
N3–C16	1.307 (3)	1.309 (3)	1.308 (3)	1.307 (4)	1.307 (3)
N4–C16	1.356 (3)	1.350 (3)	1.347 (3)	1.350 (4)	1.356 (3)
bond angles
N1–Re–N2	74.29 (7)	74.60 (8)	74.40 (8)	75.29 (11)	74.90 (7)
N1–Re–N3	80.26 (7)	82.12 (8)	83.41 (8)	79.19 (10)	79.35 (7)
N2–Re–N3	82.97 (7)	80.44 (8)	79.26 (8)	83.85 (10)	82.00 (7)
Re–N3–H3N	110 (2)	110 (2)	108 (2)	107 (3)	109 (2)
Re–N3–C16	136.66 (16)	135.62 (19)	136.74 (19)	135.4 (2)	137.12 (15)
C16–N3–H3N	113 (2)	115 (2)	115 (2)	115 (3)	114 (2)
N3–C16–N4	123.2 (2)	122.9 (2)	123.0 (2)	124.2 (3)	122.84 (19)
N3–C16–C17	119.5 (2)	119.8 (2)	119.2 (2)	118.4 (3)	119.85 (19)
N4–C16–C17	117.3 (2)	117.3 (2)	117.8 (2)	117.3 (3)	117.28 (19)
C16–N4–C18	122.62 (19)	122.4 (2)	123.2 (2)	124.2(3)	122.27(18)
C16–N4–C(n)a	122.93 (19)b	121.2 (2)c	121.0 (2)d	123.7 (3)e	121.29 (19)f

^an varies in number according to the R group.

^bn = 22.

^cn = 23.

^dn = 24.

^en = 21.

^fn = 21.

In all but one of the new complexes, the amidine ligand has a similar orientation (specified by the projection onto the equatorial plane of the amidine plane defined by the N3, C16 and N4 atoms). In this orientation, the amidine plane bisects the two N–Re–C angles in the equatorial plane. In [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂(CH₂)₂)]BF₄ (4), the amidine plane orientation is different: it is rotated by about 65°, with the methyl group almost directly above one carbonyl ligand. However, in solution there is no evidence for this difference in orientation, as the ¹H NMR signals of 4 have chemical shifts similar to those of other [Re(CO)₃(5,5′-Me₂bipy(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (3, 5, 6, 7). The different orientation in 4 is thus attributed to subtle packing effects. Furthermore, the structures of most of the complexes in this and previous studies lead us to conclude that the orientation of the amidine and iminoether ligands does not depend on the substitution pattern of the bipyridine ligands (L = 5,5′-Me₂bipy or L = 6,6′-Me₂bipy) present in the equatorial plane.^23,25

Tables 3 and 4 show that for complexes 3-12 the bond lengths from C_am (C16) to the rhenium-bound nitrogen atom (N3), and to the remote nitrogen atom (N4), are all closer to an average sp² C=N bond length (~1.28 Å), than to an average sp³ C–N bond length (~1.47 Å), as also reported for [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NHR)]BF₄ complexes²⁵ and for Ni and Cu complexes.^37–39 In addition to the C16–N3 and C16–N4 bond lengths, the values of the C16–N4–C18, C16–N4–C(n) and N3–C16–N4 angles, which are all close to 120° (Tables 3 and 4), also provide evidence for electron delocalization within the amidine group, as discussed in previous reports.^{23, 37–40} Furthermore, the N3 hydrogen atoms in these complexes are all located in positions consistent with sp² rather than sp³ hybridization for N3.

Distances that are slightly shorter for C16–N3 than for C16–N4 (Tables 3 and 4) indicate more double-bond character in the C16–N3 bond. For example, in [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ (3) the C16–N3 bond length is 1.306(4) Å and the C16–N4 bond length is 1.346(5) Å. Similar differences in the C16–N3 and C16–N4 bond distances reported previously for [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NHR)]BF₄ complexes were attributed to greater double-bond character for the C16–N3 bond than for the C16–N4 bond.²⁵

In the solid state, [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NHR)]BF₄ complexes exist as the E′ isomer.²⁵ In solutions made with polar solvents such as acetonitrile, the E′ isomer equilibrated to a mixture of the E′ and Z isomers. This equilibration, involving sequential rotations around the C16–N4 bond (fast step forming the E isomer as an undetectable intermediate in polar solvents) and then around the C16–N3 bond (slow step), required several minutes.²⁵ The solution results are consistent with the X-ray data that indicate more double-bond character in the C16–N3 bond than in the C16–N4 bond. In solvents with low polarity, such as chloroform, abundant amounts of E, E′ and Z isomers were found. Two-dimensional NMR data demonstrated that the E′ to E interconversion, involving rotation around the C16–N4 bond (Figure 1), was fast. The similarity in the C16–N3 and C16–N4 bond distances in new and old amidine complexes indicates that E to Z isomer interconversion should be slow for the new amidine complexes (3-12) as well. Thus, the NMR evidence (see below) for the presence of only one isomer on dissolution of crystals containing only the E isomer indicates beyond doubt that this one isomer is the E isomer and that the Z isomer of the new [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (3-12) is unstable. Preliminary data suggest that the rotation around the C16–N4 bond does occur and studies are planned to elucidate this process.

Steric Interaction of the Amidine Axial Ligand with the Equatorial Ligands

For amidine complexes 3-12, one of the bond angles from an equatorial N atom to the axial N3 atom (N1–Re–N3 or N2–Re–N3) is always significantly greater than the other such angle. For example, in complexes 3, 6, 7, 8, 11 and 12, the N2–Re–N3 angle is greater than the N1–Re–N3 angle, whereas in complexes 4, 5, 9 and 10 the N1–Re–N3 angle is the larger (cf. Tables 3 and 4). The smaller N–Re–N3 angle is always the one involving the equatorial N closest to the amidine N3H group. A similar relationship was also evident between the smaller equatorial N–Re–N3 bond angle and the orientation of the N3H group of previously studied primary amidine²⁵ and iminoether²³ complexes, when the axial ligand was oriented in the normal way. For the new complexes, this normal orientation is shown in Supporting Information. The reason that one N–Re–N3 bond angle is significantly larger than the other N–Re–N3 bond angle in complexes 3-12 is clearly because the larger angle leads to reduced repulsions between the amidine methyl group and the closest atoms of equatorial ligands.

When we began our investigations into reactions of coordinated acetonitrile in complexes with the fac-[M^I(CO)₃] core, one initial goal was to explore the effect of increasing the steric bulk near the metal center by using the 6,6′-Me₂bipy ligand. In the first such study (involving iminoether complexes), we found that, when the iminoether was oriented in the same way, the value of the larger N–Re–N3 angle in [Re(CO)₃(bipy)(HNC(CH₃)OCH₃)]BF₄ was greater than the corresponding larger N–Re–N3 value in [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)OCH₃)]BF₄.²³ We hypothesized that the distortion in the 6,6′-Me₂bipy complex decreases those interactions of the axial iminoether ligand with the equatorial ligands that cause one of the two N–Re–N3 angles to be larger.

In the new complexes, the size of the larger of the two N–Re–N3 bond angles in the 5,5′-Me₂bipy complexes is greater on average than the larger bond angles in the 6,6′-Me₂bipy complexes (Tables 3 and 4 and Supporting Information). This comparison supports the hypothesis that the distortion in the 6,6′-Me₂bipy complexes decreases those axial-equatorial ligand interactions that cause one of the two N–Re–N3 angles to be larger. This apparently counter-intuitive finding of smaller interactions in 6,6′-Me₂bipy complexes than in the related [Re(CO)₃(bipy)(HNC(CH₃)OCH₃)]BF₄ complexes can be understood by considering our structural results and those that have appeared during the course of our work.⁴¹ In the many structures now available, the clashes between the methyl groups of the 6,6′-Me₂bipy ligand and the two equatorial CO ligands distort the 6,6′-Me₂bipy ligand and force the 6,6′-methyl groups out of the equatorial plane (defined by the C13–Re–C14 atoms) toward the axial CO. These distortions of the Re(CO)₃(6,6′-Me₂bipy) moiety in [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ amidine complexes (8-12) (Figure 5 and Supporting Information) are very similar to those of the other complexes.^23,41

Figure 5.

Views of piperidinylamidine complexes, [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄, depicted with the C13–Re–C14 equatorial plane perpendicular to the plane of the paper. Shown at left and middle are front and side views, respectively, of complex 8 with L = 6,6′-Me₂bipy. Pictured at right is a side view of complex 3 with L = 5,5′-Me₂bipy.

As can be seen in Figure 5, the distortion results in a tilted plane of the 6,6′-Me₂bipy ligand. To appreciate the effect of the tilting, it is convenient to view the two Me₂bipy ligands as having an interior or front side (atoms N1, C1, N2, C10) and an exterior or back side (atoms C3, C4, C7, C8), according to the numbering scheme in Figures 3 and 4. Although in the solid state the ligands are not fully symmetrical or fully planar, the front-side carbons 1 and 10 lie slightly below the equatorial plane in 6,6′-Me₂bipy complexes and lie in the equatorial plane in 5,5′-Me₂bipy complexes. To assess the space near the axial coordination site (trans to the axial CO), we measured some non-bonded distances from N3 (Supporting Information). For [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ (Y = CH₂ or NH), the non-bonded distances from N3 to C1 and C10 average ~0.15 Å longer in 6,6′-Me₂bipy than in 5,5′-Me₂bipy complexes. Properties (such as N–Re–N bond angles) affected by the interior structure have values (Tables 3 and 4) consistent with this additional space. On the other hand, for these same complexes the non-bonded distances from N3 to C4 and C7 average ~0.5 Å shorter in 6,6′-Me₂bipy than in 5,5′-Me₂bipy complexes. Other properties, such as some NMR shifts, are affected more by the exterior or peripheral structure (see below).

Furthermore, for some properties, the net effects of the differences in the bidentate ligand orientations may cancel. Indeed, regardless of whether the complex has L = 6,6′-Me₂bipy or 5,5′-Me₂bipy, the isomer distribution seems to be unaffected. Thus, for all the complexes in the present study, the repulsions are large enough to favor the presence of only one isomer, namely the E isomer.

Our ranking of the expected effects of steric interactions on isomer stability for [Re(CO)₃(Me₂bipy)(HNC(CH₃)OCH₃)]⁺, [Re(CO)₃(L)HNC(CH₃)NHR)]⁺, and [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]⁺ complexes is shown in Figure 6. This ranking summarizes our experimental observations of the relative isomer abundance of these complexes in this and previous studies.^23,25 This ranking takes into account steric repulsions of the substituents on C_am or C_ie in the axial with the equatorial Me₂bipy and CO ligands and also the relative repulsions within the amidine ligand between the CH₃ and the NH or NR groups in [Re(CO)₃(L)HNC(CH₃)NHR)]⁺ complexes. As indicated for the two structures sketched at the far right of Figure 6, the repulsive interactions of the NR substituent with the equatorial ligands depicted in the respective Z′ and Z sketches are expected to be the most severe. Thus, these interactions are shown with thicker double-headed arrows. The order of the Z and E isomers of [Re(CO)₃(L)HNC(CH₃)NHR)]⁺ complexes (4th and 5^th structures from left in Figure 6) reflects our suggestion that the N4H interaction with the equatorial ligands is less repulsive than the corresponding C_amCH₃ interaction with the equatorial ligands.

Figure 6.

Ranking of increasingly unfavorable total steric repulsive interactions (each double-headed arrow indicates an interaction) in [Re(CO)₃(Me₂bipy)(HNC(CH₃)OCH₃)]⁺, [Re(CO)₃(L)HNC(CH₃)NHR)]⁺ and [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]⁺ complexes [N–N, denotes the 5,5′- or 6,6′-Me₂bipy ligands, and Y = CH₂, (CH₂)₂, (CH₂)₃, NH, or O].

Repulsion between the CH₃ and NR groups is secondary and noticeably influences abundance mainly when the two isomers have the same interaction with the equatorial ligands, such as is the case with the E′ and E isomers of [Re(CO)₃(L)HNC(CH₃)NHR)]⁺ complexes (3rd and 5^th structures from left in Figure 6). For [Re(CO)₃(L)HNC(CH₃)NHR)]⁺ complexes, 25 clashes between the NR and the CH₃ amidine substituents destabilize the E isomer, which normally has low abundance. The abundance of the E′ isomer increased as the steric bulk of the R substituent on C_am increased. In turn, the Z isomer of fac-[Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NH₂)]BF₄ with similarly sized substituents (NH₂ and CH₃) on C_am was highly favored (~ 90% abundant). We caution that the differences in electronic effects influencing the stability of the Z and E configurations are not known. Nevertheless, the ranking as illustrated in Figure 6 does provide a good guide for predicting the relative abundance of the isomers, especially in polar solvents.

NMR Spectroscopy

All complexes were characterized by ¹H NMR spectroscopy in acetonitrile-d₃; selected complexes were also studied in CDCl₃ and DMSO-d₆. ¹H NMR spectra were recorded within at least 6 min of dissolution. In contrast to the spectral data of the previously studied [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NHR)]BF₄ complexes, all of the ¹H NMR spectra of the new [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ amidine complexes regardless of the solvent used consistently indicate the presence of only one isomer in solution. Moreover, the spectra of all of these complexes (3-12) showed no changes with time, even after several days.

The atom numbering system used in this NMR discussion is that shown in Figures 3 and 4. ¹H NMR signals of the bidentate ligand and of N3H were assigned by using the splitting pattern and integration, and by comparison to unambiguous assignments of spectra for previously reported analogous Re^I amidines and iminoether complexes.^23,25

We illustrate our findings by detailing our studies of compound 3, [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄. When crystals of 3 were dissolved in three different solvents (acetonitrile-d_3, CDCl₃ and DMSO-d₆), ¹H NMR spectra showed no evidence for more than one isomer: All peaks in all three solvents remained constant when solutions were monitored from 3 min after dissolution until two weeks. As indicated in our analysis of the C16–N3 bond lengths above, we believe that if the Z isomer were present, the interconversion rate would be slow and we would have detected signals for the Z isomer. Thus, we are absolutely confident that the Z isomer is unstable.

The N3H signal in the new complexes is easily assigned because the peak is a broad singlet integrating to one proton and because it disappeared gradually after the addition of D₂O. For 3, this N3H signal has a more downfield shift in DMSO-d₆ (5.77 ppm) than in acetonitrile-d₃ (4.78 ppm) or CDCl₃ (4.60 ppm). The related values for 4 were 5.32, 4.52, and 4.32 ppm, respectively. A similar NMR dependence of the N3H shift on solvent was observed for [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)OCH₃)]BF₄; in a standard chloride titration experiment, the downfield shift in DMSO-d₆ was demonstrated to be caused by hydrogen bonding of N3H to DMSO-d₆.23 In this iminoether complex, as for complexes 3 and 4, N3H projects out toward the solvent, making this proton available for hydrogen bonding to DMSO-d₆. Such hydrogen bonding explains the solvent dependence found for 3 and 4.

Dependence on Y of the N3H NMR Signals of [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄, for L = 5,5′-Me₂bipy and 6,6′-Me₂bipy

Selected ¹H NMR signals of the new amidine complexes (3-12) in acetonitrile-d₃ are compared in Table 5. For complexes with the amidines having six-membered N(CH₂CH₂)₂Y rings, [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ (3, 6, and 7) and [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ (8, 11, and 12), the most downfield shift observed for the N3H signal is for the morpholine derivative (Y = O) in each series (4.94 ppm for 7 and 5.30 ppm for 12). The N3H signal is slightly upfield for piperazine derivatives (Y = NH) (4.84 ppm for 6 and 5.18 ppm for 11) and farther upfield for piperidine derivatives (Y = CH₂) (4.78 ppm for 3 and 5.14 ppm for 8). These data indicate that the remote O and N atoms of the morpholine and piperazine derivatives, respectively, exert electron-withdrawing effects on the amidine group, with the more electronegative O atom of the morpholine derivative having the greater downfield-shifting effect on the N3H signal. In the two series, the N3H signal systematically shifted upfield as the size of the ring increased from six to seven to eight members. The most upfield N3H shift observed was for the heptamethyleneimine derivatives (Y = (CH₂)₃) with the eight-membered ring. The variations in NH shift as the ring size changes can be attributed to a combination of ring-strain, inductive, and solvation effects.

Table 5.

¹H NMR Shifts (ppm) for L, N3H, and C_amCH₃ in Complexes (Acetonitrile-[Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ Complexes (Acetonitrile-d₃, 25 °C)

Y	H3/3′	H4/4′	H5/5′	H6/6′	L-CH3	N3H	CamCH3
L = 5,5′-Me2bipy
CH2 (3)	8.26	8.04		8.85	2.48	4.78	2.10
(CH2)2 (4)	8.27	8.05		8.87	2.48	4.52	2.10
(CH2)3 (5)	8.27	8.04		8.87	2.48	4.49	2.12
NH (6)	8.26	8.04		8.85	2.47	4.84	2.12
O (7)	8.26	8.04		8.85	2.48	4.94	2.14
L = 6,6′-Me2bipy
CH2 (8)	8.19	8.06	7.62		3.06	5.14	1.60
(CH2)2 (9)	8.19	8.06	7.61		3.07	4.90	1.62
(CH2)3 (10)	8.19	8.07	7.63		3.07	4.82	1.66
NH (11)	8.19	8.07	7.62		3.05	5.18	1.63
O (12)	8.19	8.07	7.62		3.05	5.30	1.66

Dependence on L of the C_amCH₃ NMR Signals of [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄, for L = 5,5′-Me₂bipy and 6,6′-Me₂bipy

We can readily explain the differences in ¹H NMR shifts of the C_amCH₃ signal between the two series, [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ (~2.1 ppm, 3-7) and [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ (~1.6 ppm, 8-12). The shifts are very similar within each of the two series (Table 5). The more upfield shift (by ~0.5 ppm) of the C_amCH₃ signal for the 6,6′-Me₂bipy complexes (8-12) than for the 5,5′-Me₂bipy complexes (3-7) is clearly attributable to the anisotropic effect of the 6,6′-Me₂bipy aromatic ring system. Compared to the [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (3-7), all of the Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (8-12) have a shorter distance from the methyl carbon of the amidine ligand (C17) to the centroid of the closest bipyridine ring. This shorter distance results from the tilting in the 6,6′-Me₂bipy ligand, moving the back side of the ring up toward the amidine as discussed above. For example, these distances in [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)NC₅H₁₀)]BF₄ (3) and [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)NC₅H₁₀)]BF₄ (8) are 4.2 Å and 3.4 Å, respectively. Therefore, the anisotropic upfield-shifting effect of the bipyridine rings is greater on the C_amCH₃ methyl signal of [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes 8-12 than on the C_amCH₃ ¹H NMR signal for [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes 3-7.

The N3H shifts of [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes for L = 6,6′-Me₂bipy are downfield from the corresponding shifts of the L = 5,5′-Me₂bipy analogues (Table 5). At this time, we cannot identify the reasons for this difference because as mentioned above, N3H shifts are influenced by a multiplicity of possible factors. In addition, as L is changed, any changes in the heavy-atom anisotropic or inductive effects of the Re will affect the shift.

Dependence of Reaction Times on the Me₂bipy Ligand and the Amine

For a given amine, reactions were relatively faster with [Re(CO)₃(6,6′-Me₂bipy)(CH₃CN)]BF₄ (2) than with [Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]BF₄ (1) (Table 6). For 1 and 2, the time required for complete reaction, assessed by checking for reaction completion from time to time by NMR spectroscopy (Figure 7), varied with basicity and the ring size of the heterocyclic amine. The pK_a values of the heterocyclic amines⁴² (Table 6) decrease in the order, piperidine (with the highest pK_a, 11.1) > homopiperidine > heptamethyleneimine > piperazine > morpholine.⁴² The reactions of [Re(CO)₃(6,6′-Me₂bipy)(CH₃CN)]BF₄ (2) with piperidine and piperazine were essentially complete before the first spectrum could be recorded (≤ 3 min). Morpholine, the other six-membered-ring amine, required a much longer reaction time (30 min) owing to its lower basicity (pK_a = 8.5). This same pattern as found for 2 was observed with these heterocyclic amines for 1. For example, morpholine had the longest reaction completion time (4 h) for the six-membered ring amines with 1 (Figure 7). These results indicate that greater heterocyclic amine basicity is associated with faster reactions, as would be expected. Piperazine has the second lowest pK_a (10.2) compared to the other heterocyclic amines used here; however, the reactions of piperazine with 1 and 2 were relatively fast (≤ 3 min and 20 min, respectively). This relative reactivity can be attributed to the statistical reaction probability for each piperazine molecule (with two amine groups) being twice that of other amines used.

Table 6.

Times for Complete Reaction of [Re(CO)₃(L)(CH₃CN)]BF₄ Complexes with Heterocyclic Amines (HN(CH₂CH₂)₂Y) to Form [Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ Complexes^a

HN(CH2CH2)2Y (Y)	pKa	L = 6,6′-Me2bipy	L = 5,5′-Me2bipy
piperidine (CH2)	11.1	≤3 min	<5 min
homopiperidine ((CH2)2)	10.9	~4.5 min	~8 min
heptamethyleneimine ((CH2)3)	10.8	~6 min	6 h
piperazine (NH)	10.2	≤3 min	20 min
morpholine (O)	8.5	30 min	4 h

^aReaction monitored by NMR spectroscopy in acetonitrile-d₃ at 25 °C.

Figure 7.

Aromatic region of the ¹H NMR spectra in acetonitrile at 25 °C of the reaction of [Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]BF₄ (1) with morpholine to form [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂O)]BF₄ (7).

A comparison of reaction completion times for amines with no other heteroatoms in the ring (Table 6) is instructive. For both the [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ and the [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ series, the reaction times increase in the order, piperidine < homopiperidine < heptamethyleneimine (Table 6). This finding of longer reaction completion times as the number of amine methylene groups increases makes it clear that steric effects decrease amine reactivity. However, the effect of amine bulk on reaction time is highly pronounced only for heptamethyleneimine with the 5,5′-Me₂bipy complex 1. The effect is much less pronounced for the 6,6′-Me₂bipy complex 2 because of the greater interior space near the axial coordination site caused by the tilting of the 6,6′-Me₂bipy ligand (as described above).

Reactions of most cyclic secondary amines with [Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]BF₄ (1) reached completion in less than 1 h (Table 6). In contrast, more time was required for reactions of 1 with primary aliphatic amines, even though most of these previously studied amines have a basicity lying within the pK_a range in Table 6. For example, the reaction of [Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]BF₄ (1) required ~6 h for methylamine (pK_a⁴³ = 10.6) and ~4 days for tert-butylamine (pK_a⁴³ = 10.5).²⁵ Reaction times of [Re(CO)₃(5,5′-Me₂bipy)(CH₃CN)]BF₄ (1) and [Re(CO)₃(6,6′-Me₂bipy)(CH₃CN)]BF₄ (2) with isopropylamine (pK_a⁴³ = 10.6) are 28 h and 14 h, respectively.²⁴ These results are consistent with the expected lower nucleophilicity of primary amines as compared with that of the cyclic secondary amines studied here.

Robustness of the Piperidinylamidine Ligation

A fivefold excess of the relatively basic, strongly coordinating 4-dimethylaminopyridine ligand was added to [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ (3) in acetonitrile-d₃ or in CDCl₃. No changes in spectral features of 3 were observed for up to two months, indicating that the piperidinylamidine ligand is not readily replaced. The NMR signals for [Re(CO)₃(5,5′-Me₂bipy)(4-dimethylaminopyridine)]BF₄,25 synthesized as a control, did not change with time in either acetonitrile-d₃ or CDCl₃.

Conclusions

Unlike previously studied analogous amidine complexes derived from primary amines, all ten of the [Re(CO)₃(5,5′- or 6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes formed from cyclic secondary amines studied here exist as only one isomer (the E isomer) in both the solid state and in solution. These findings are attributable to the combination of the high steric bulk and the C₂ symmetry of the amidine substituents. After dissolution and sufficient time for equilibrium to be established in solution, only the initial E isomer was detectable. Thus, the equilibrium between the Z and E isomers must lie far to the side of the E isomer. We conclude that steric repulsions between the N(CH₂CH₂)₂Y groups of the axial amidine ligands and the equatorial ligands preclude formation of any isomer other than E (Figures 2 and 6). Nevertheless, these repulsive interactions do not lead to a weakened Re–N3 bond, as indicated by the length of this bond.

The 6,6′-methyl groups in [Re(CO)₃(6,6′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄ complexes (8-12) cause the 6,6′-Me₂bipy ligand to distort and tilt. Although the “front side” of the 6,6′-Me₂bipy ligand with the 6,6′-methyl groups projects down toward the axial CO group, the “back side” of the 6,6′-Me₂bipy ligand projects up. Thus the 6,6′-Me₂bipy ligand has a net steric footprint comparable to that of the untilted 5,5′-Me₂bipy ligand.

The [Re(CO)₃(5,5′-Me₂bipy)(HNC(CH₃)N(CH₂CH₂)₂CH₂)]BF₄ complex (3) in acetonitrile-d₃ or in CDCl₃ was robust when challenged with 4-dimethylaminopyridine, indicating that amidine ligands are strong donors. The heterocyclic amines employed here have a relatively high reactivity and form only amidines with the E configuration, indicating that amidine complexes can be formed quickly and isomerically pure. All of these favorable properties cited here suggest that the strategy of using heterocyclic amines to create amidine links to the fac-[M(CO)₃]⁺ core (M = ^99mTc and ^186/188Re radionuclides) may be a useful conjugation method for the development of targeted radiopharmaceuticals.

Synopsis.

SynopsisDiverse C₂-symmetrical saturated heterocyclic amines, HN(CH₂CH₂)₂Y, rapidly add to fac-[Re(CO)₃(L)(CH₃CN)]BF₄ (L = dimethyl-2,2′-bipyridines). The new amidine complexes, fac-[Re(CO)₃(L)(HNC(CH₃)N(CH₂CH₂)₂Y)]BF₄, have the novel feature of existing as only one isomer (amidine E configuration). The rapid reaction, the lack of isomers, and the robust attachment of the amidine ligand offer promise that the amidine group could serve as a linking group in the conjugation of the fac-[M(CO)₃]⁺ core (M = ^99mTc and ^186/188Re radionuclides) with biomedical targeting moieties.