Bis-Monophospholyl Dysprosium Cation Showing Magnetic Hysteresis at 48 K.

Single-molecule magnets (SMMs) have potential applications in high-density data storage, but magnetic relaxation times at elevated temperatures must be increased to make them practically useful. Bis-cyclopentadienyl lanthanide sandwich complexes have emerged as the leading candidates for SMMs that show magnetic memory at liquid nitrogen temperatures, but the relaxation mechanisms mediated by aromatic C5 rings have not been fully established. Here we synthesize a bis-monophospholyl dysprosium SMM [Dy(Dtp)2][Al{OC(CF3)3}4] (1, Dtp = {P(CtBuCMe)2}) by the treatment of in-situ-prepared "[Dy(Dtp)2(C3H5)]" with [HNEt3][Al{OC(CF3)3}4]. SQUID magnetometry reveals that 1 has an effective barrier to magnetization reversal of 1760 K (1223 cm-1) and magnetic hysteresis up to 48 K. Ab initio calculation of the spin dynamics reveals that transitions out of the ground state are slower in 1 than in the first reported dysprosocenium SMM, [Dy(Cpttt)2][B(C6F5)4] (Cpttt = C5H2tBu3-1,2,4); however, relaxation is faster in 1 overall due to the compression of electronic energies and to vibrational modes being brought on-resonance by the chemical and structural changes introduced by the bis-Dtp framework. With the preparation and analysis of 1, we are thus able to further refine our understanding of relaxation processes operating in bis-C5/C4P sandwich lanthanide SMMs, which is the necessary first step toward rationally achieving higher magnetic blocking temperatures in these systems in the future.

Introduction

The potential for high-density data storage devices based on single-molecule magnets (SMMs) is reliant upon increasing spin relaxation times toward practically useful time scales at relatively high temperatures, away from expensive liquid helium regimes to that of cheap and abundant liquid nitrogen.[1] Lanthanide (Ln) based SMMs have been at the forefront of research in this area for the past 15 years,[2] and design principles popularized by Rinehart and Long in 2011[3] directed the community toward longer relaxation times by means of massive increases in the energy barrier to magnetic reversal (Ueff).[4] These large increases in Ueff did not lead to corresponding increases in magnetic remanance temperatures[5] until the dysprosocenium cation [Dy(Cpttt)2]+ (Cpttt = C5H2tBu3-1,2,4) was shown to exhibit magnetic hysteresis at TH = 60 K in 2017.[6] We attributed the high-temperature magnetic remanance in this bis-Cpttt system to the combination of a Dy3+ center with rigid, charge-dense π-aromatic rings; we also predicted that removal of C–H groups from the C5 ring could increase hysteresis temperatures further.[6a] This has been proven correct, with hysteresis temperatures up to TH = 80 K observed for peralkylated bis-cyclopentadienyl Ln complexes reported in the past two years.[7]

An alternative strategy to remove C–H groups from C5 frameworks is heteroatom substitution;[8] indeed, the magnetic properties of theoretical [Dy(E5)2]+ (E = N, P) cations have recently been predicted to exhibit high Ueff values.[9] Phospholyl ligands are a valid alternative to cyclopentadienyls as the P lone pairs are relatively soft, so they tend to exhibit η5-binding modes with Ln ions.[10] Synthetic routes toward peralkylated monophospholyls are already mature; of most relevance here, the Ln chemistry of {P(CtBuCMe)2} (Dtp) has already been developed.[11] The straightforward synthesis of [Dy(Dtp)2(I)] from DyI3 and 2 equiv of KDtp was reported by Nief and co-workers in 2009,[11d] and we envisaged that this would be an ideal starting material toward the isolation of a [Dy(Dtp)2]+ cation. Herein we report the synthesis and magnetic properties of this cation and correlate our results with ab initio calculations of the spin dynamics to probe the effects of aromatic P–C vibrational modes in magnetic relaxation mechanisms compared to aromatic C–C vibrations. We find that relaxation is expedited in the [Dy(Dtp)2]+ cation compared to [Dy(Cpttt)2]+ as additional vibrational modes are brought on-resonance, providing new insights into the relaxation pathways that operate in bis-C4P vs bis-C5 Ln sandwich SMMs.

Results and Discussion

Synthesis

Treatment of “[Dy(Dtp)2(C3H5)]” with [NEt3H][Al{OC(CF3)3}4] in benzene gave [Dy(Dtp)2][Al{OC(CF3)3}4] (1) following workup and recrystallization from chlorobenzene (Scheme ). “[Dy(Dtp)2(C3H5)]” was prepared from the salt metathesis reaction of [Dy(Dtp)2(I)][11d] with C3H5MgCl, while [NEt3H][Al{OC(CF3)3}4] was isolated from the reaction of Li[Al{OC(CF3)3}4][12] with NEt3HCl by adapting procedures used for the synthesis of [NEt3H][B(C6F5)4].[13] Crude “[Dy(Dtp)2(C3H5)]” was obtained as an orange foam and was used in situ without further purification; we were unable to isolate the pure complex to collect meaningful characterization data due to its high solubility in pentane, but we are confident of its formulation from the formation of 1. The global yield of 1 is 26% over three reaction steps based on DyI3. The triethylammonium reagent was selected as it provides an entropic driving force with dual amine and alkene elimination during the reaction,[14] and the [Al{OC(CF3)3}4]− anion is more weakly coordinating than the [B(C6F5)4]− anion,[15] which has been used for the synthesis of all Ln metallocenium cations to date.[6,7,16] The direct reaction of [Dy(Dtp)2(I)] with [H(SiEt3)2][B(C6F5)4][17] gave an intractable mixture of products. 1H, 13C, and 31P NMR spectra of a sample of 1 in d5-chlorobenzene were uninformative due to paramagnetism, but the [Al{OC(CF3)3}4]− anion was detected by 19F NMR spectroscopy (δF: −90.50 ppm; v1/2 = 300 Hz); the presence of paramagnetic [Dy(Dtp)2]+ cations has broadened this signal and shifted it considerably from the [NEt3H][Al{OC(CF3)3}4] precursor (δF: −75.70 ppm, d2-DCM).

Scheme 1

Synthesis of 1

Structural Characterization

The solid-state structure of 1 was determined by single-crystal X-ray diffraction (Figure ). The [Dy(Dtp)2]+ cation in 1 exhibits a bent geometry, with a Dtpcent···Dy···Dtpcent angle of 157.94(4)° and mean Dy···Dtpcent distances of 2.354(3) Å; although this is slightly less bent than [Dy(Cpttt)2][B(C6F5)4] (Cptttcent···Dy···Cptttcent, 152.56(7)°; Dy···Cptttcent, 2.316(3) Å),[6a] the incorporation of phosphorus in the rings has led to increased Dy–ligand distances. As expected from removal of an equatorial iodide, the Dy3+ center in 1 has a larger Dtpcent···Dy···Dtpcent angle and shorter Dy···Dtpcent distances than the precursor [Dy(Dtp)2(I)] (Dtpcent···Dy···Dtpcent, 147.29(3)°; mean Dy···Dtpcent, 2.416(2) Å).[11d] The Dtp rings in 1 are staggered with respect to each other, with the phosphorus atoms at adjacent positions (mean Dy–P, 2.7931(11) Å). Although an η5-binding mode is adopted, there is a significant variation in Dy–CDtp distances: range 2.570(3)–2.780(3) Å, cf., 2.702(2)–2.778(2) Å for [Dy(Dtp)2(I)].[11d] The electron deficient Dy3+ center in 1 forms additional stabilizing electrostatic contacts with tBu groups, with two short Dy···C (2.881(3) and 3.026(4) Å) and two short Dy···H (2.481 and 2.541 Å) distances; similar metrical parameters for electrostatic interactions between Dy3+ centers and two C–H bonds of tBu groups were previously seen for [Dy(Cpttt)2][B(C6F5)4] (Dy···C, 2.964(5) Å mean; Dy···H, 2.4989 Å mean).[6a] The metrical parameters of the [Al{OC(CF3)3}4]− anions are unremarkable, and these do not show any interaction with the Dy3+ center (shortest Dy···F distance >6.0 Å).

Figure 1

Molecular structure of the cation of [Dy(Dtp)2][Al{OC(CF3)3}4] (1) with selected atom labeling: left, side view; right, top view. Displacement ellipsoids are set at the 50% probability level, and the anion and hydrogen atoms are omitted for clarity. Dy atoms are teal. P atoms are purple, and C atoms are gray. Selected bond distances (Å) and angles (deg): Dy(1)···Dtpcent(1), 2.355(2); Dy(1)···Dtpcent(2), 2.352(2); Dy(1)···P(1), 2.7981(8); Dy(1)···P(2), 2.7880(8); range Dy(1)···CDtp, 2.570(3)–2.780(3); Dy(1)···C(8), 2.881(3); Dy(1)···C(22), 3.026(4); Dy(1)···H(8A), 2.481; Dy(1)···H(22A), 2.541; Dtpcent(1)···Dy(1)···Dtpcent(2), 157.94(4); C(8)···Dy(1)···C(22), 115.38(9).

Magnetism

A polycrystalline sample of 1 suspended in eicosane was analyzed by SQUID magnetometry to determine its magnetic properties. The magnetic susceptibility temperature product (χMT) of solid 1 at 300 K is 13.85 cm3 K mol–1 (Supporting Information Figure S8); this is in accord with the free-ion Curie value of 14.17 cm3 K mol–1[18] and ab initio calculations (13.66 cm3 K mol–1, see below). A steady reduction in χMT with temperature for solid 1 was observed down to 25 K (12.28 cm3 K mol–1), owing to thermal depopulation of the excited crystal field (CF) states. A more severe drop in χMT was observed below 25 K due to the onset of magnetic blocking, which correlates with the temperature at which the zero-field cooled (ZFC) susceptibility has a plateau (TB1 = 25 K; Figure S10). The nontraditional profiles of the field cooled (FC) and ZFC susceptibilities are a complicated function of the measurement protocol (temperature sweep rate, magnetic field strength, and field sweep rate) as well as the intricate field and temperature dependence of magnetic relaxation in Dy3+ SMMs;[19] such traces have been explained by others.[20] The most salient information from the FC/ZFC traces is the temperature at which the two data sets bifurcate: for 1, Tirrev = 54 K (Figure S11).

Slow relaxation of magnetization for 1 was confirmed by the presence of out-of-phase maxima between 60 and 80 K in the zero-field ac susceptibility data (Figures S12 and S13). The temperature dependence of the relaxation times obtained from these measurements were fitted to a generalized Debye model using CC-FIT2[21] (Figure ), which allows the extraction of uncertainties in the magnetic relaxation times from the underlying distribution function. We observe an exponential relaxation process (Orbach mechanism; τ–1 = τ0–1 exp[−Ueff/T]) above 50 K and extract an effective barrier to magnetization reversal Ueff = 1760(70) K (1220(50) cm–1), with τ0 = 10–11.7(4) s (ca. 2 × 10–12 s). The Ueff value for 1 is identical to that previously seen for [Dy(Cpttt)2][B(C6F5)4] (1760 K),[6a] and smaller than the current record-holder [Dy(C5iPr5)(C5Me5)][B(C6F5)4] (2217 K).[7b] To obtain relaxation times at lower temperatures, we performed magnetization decay experiments and fitted the data with stretched exponentials (Figure S14 and Table S3). Following a similar approach for obtaining uncertainties from ac data,[21] we determined uncertainties from the magnetization decay experiments based on the well-known distribution underlying the stretched exponential function (see Supporting Information, Figure S14);[22] this gives at the 1σ level, where β is the stretch factor. Below 30 K we observe a power-law relaxation process (Raman-like mechanism; τ–1 = CT) for 1, and this data is well-reproduced with n = 1.1(3) and C = 10–3.5(3) s–1 K– (ca. 3 × 10–4 s–1 K–). The small n value approaches that expected for the direct relaxation process;[23] however, as these data are collected in zero magnetic field, this is not a plausible mechanism. Indeed, all bis-cyclopentadienyl Dy3+ cations have relatively low Raman exponents of between 2 and 3 in the crystalline phase,[6,7] and thus, substitution of C for P in the first coordination sphere of 1 does not appear to grossly alter this characteristic; however, it cannot be ascertained if the even lower exponent of 1.1(3) here is due to the effect of the ring substitution or to the different counterion ([Al{OC(CF3)3}4]−, cf., [B(C6F5)4]− for all dysprosocenium SMMs to date). While we cannot measure the relaxation dynamics between 30 and 64 K, extrapolation of the Orbach and Raman regions suggests that they intersect at 52 K which coincides with the bifurcation of FC/ZFC plots (Tirrev = 54 K): such a sharp intersection between the Raman and Orbach regions was observed for [Dy(Cpttt)2][B(C6F5)4], as was the coincidence of the intersection temperature and Tirrev.[6a] Using magnetization decays we have been able to directly measure the 100 s blocking temperature as TB2 = 23 K. Overall, magnetic relaxation is around 10–100 times faster in the range 2–100 K for 1 than for [Dy(Cpttt)2][B(C6F5)4] (Figure S15).

Figure 2

Temperature dependence of the magnetic relaxation rate of 1. Red circles are the relaxation rates extracted from ac susceptibility data (high temperature) and dc magnetization decay data (low temperature); solid red lines are error bars from the distributions of relaxation times (see Supporting Information).[21] The solid blue line is given by τ–1 = τ0–1 exp[−Ueff/T] + CT. The dashed green line is given by τ–1 = τ0–1 exp[−Ueff/T], and the dotted orange line is given by τ–1 = CT with Ueff = 1760(70) K, τ0 = 10–11.7(4) s, C = 10–3.5(3) s–1 K–, and n = 1.1(3).

Solid 1 shows open, but comparatively waist-restricted, magnetic hysteresis loops up to TH = 48 K (Figure ), using a sweep rate of ca. 20 Oe/s around the important zero-field region where quantum tunneling of the magnetization (QTM) dominates for Ln SMMs.[4] The value of TH for 1 is lower than the majority of isolated dysprosocenium cations reported to date, which have shown TH values of 60–80 K,[6,7] except for one example, [Dy(C5iPr4H)2][B(C6F5)] (TH = 32 K),[7a] which contains ring C–H protons that have been postulated to enhance magnetic relaxation mechanisms.[6a] Despite the lack of ring protons in 1, it shows open hysteresis to a maximum temperature that is 12 K lower than that previously seen for [Dy(Cpttt)2][B(C6F5)4] (TH = 60 K).[6]

Figure 3

Magnetic hysteresis of solid 1, measured with a mean field sweep rate of 21(9) Oe s–1 for |H| < 10 kOe, 49(12) Oe s–1 for 10 < |H| < 20 kOe, and 88(17) Oe s–1 for 20 < |H| < 70 kOe. Hysteresis loops recorded from 2 to 18 K in 2 K steps, from 20 to 40 K in 5 K steps, and from 43 to 50 K in 1 K steps.

Ab Initio Calculations

First-principles complete active space self-consistent field spin–orbit (CASSCF-SO) calculations were performed on the crystal structure of the cation in 1 to complement experimental data and to probe magnetic relaxation mechanisms (Table S4). As expected for a strongly axial CF, we observe an easy-axis ground Kramers doublet corresponding to the m = ±15/2 CF state, where the first five excited states are also easy-axis-like and collinear with the ground doublet; the five excited states are dominated by m = ±13/2, ±11/2, ±9/2, ±7/2, and ±5/2, respectively. The g-values for the sixth excited Kramers doublet are highly rhombic, indicating a substantially mixed m composition, and thus magnetic relaxation by the Orbach process is likely to occur via this state (ca. 1716 K, which compares reasonably well with the experimental Ueff = 1760(70) K). To gain more insight into the relaxation dynamics, we have calculated the spin dynamics using our previously described ab initio method;[6a] briefly, this entails the following: (i) optimization of molecular geometry and determination of vibrational modes with DFT, (ii) calculation of spin–phonon coupling with CASSCF-SO, and (iii) simulation of magnetic relaxation via a semiclassical master equation (see Supporting Information for details). We find excellent agreement with the experimental data in the high-temperature region corresponding to Orbach relaxation (Figure a); note that relaxation via two-phonon Raman processes at low temperatures is not accounted for in these calculations. Examining the calculated relaxation rates carefully, we observe that relaxation shows two different exponential processes in different temperature regimes (Figure S17), and that this has a slight dependence upon the choice of phonon line width (Table S7). We find that magnetic relaxation follows an Orbach process over an effective barrier of ca. 1600–1700 K following the pathway shown in Figure b, but at temperatures less than ca. 52 K the effective barrier is reduced to ca. 660–960 K (Table S7 and Figures S17 and S18). The experimental data for 1 show only one Orbach process with Ueff = 1760(70) K down to 64 K and the onset of Raman relaxation below 30 K. Thus, a potential crossover to a smaller Ueff regime may occur between 64 and 30 K; however, we cannot probe these time scales with our instrumentation.

Figure 4

(a) Ab initio calculated magnetic relaxation rates for 1 (lines) compared with the experimental data (points). (b) Energy barrier to magnetic relaxation for 1, calculated at 100 K and using a phonon linewidth of 6 cm–1. Electronic states from CASSCF-SO calculations, decomposed in the J = 15/2 basis. The opacity of the arrows is proportional to the single-phonon transition probability normalized from each departing state and commencing with unit population in |−15/2⟩; only relaxation pathways toward |+15/2⟩ are shown. ⟨J⟩ is the expectation value of the J operator along the quantization axis.

Decomposing the relaxation rates for the large Ueff process, the first step in magnetic relaxation is delicately balanced between the |±15/2⟩ to |±13/2⟩ and the |±15/2⟩ to |±11/2⟩ transitions: lower temperatures and larger phonon line widths favor the former, while higher temperatures and smaller phonon line widths favor the latter (Figure b, cf., Figure S19). The |±15/2⟩ to |±13/2⟩ transition is mostly driven by mode 61, which is an in-plane deformation of the rings (Figure S20), whereas the |±15/2⟩ to |±11/2⟩ transition is driven by modes 76 and 77, which involve in-phase and out-of-phase deformations of the Dy3+ center via the carbon atoms of the Dtp rings (Figure S21).

Experimentally we observe that 1 relaxes faster than [Dy(Cpttt)2][B(C6F5)4] in the Orbach regime (Figure S15), and this is also borne out in comparable simulations (Figure S22; note that we have repeated calculations for [Dy(Cpttt)2][B(C6F5)4] using the slightly revised methodology employed here, see Table S6 and Figure S16). Seemingly in contradiction with the overall calculated relaxation rates (Figure S22), the escape rate of the |±15/2⟩ state in [Dy(Cpttt)2][B(C6F5)4] between 50 and 300 K is approximately an order of magnitude faster than that for 1 (Table S8), owing to the much faster |±15/2⟩ to |±13/2⟩ transitions in [Dy(Cpttt)2][B(C6F5)4] (Table S9). However, we note that all electronic states in 1 are compressed in energy, cf., [Dy(Cpttt)2][B(C6F5)4] (Figure S23), due to a weaker crystal field, and that this brings the subsequent steps in relaxation (|±13/2⟩ to |±11/2⟩ at 229 cm–1, |±11/2⟩ to |±9/2⟩ at 151 cm–1, and |±9/2⟩ to |±7/2⟩ at 130 cm–1 for 1) into resonance with vibrational modes with significant spin–phonon coupling (Figures S24 and S25 and Table S10). These excitations are 267, 172, and 157 cm–1 for [Dy(Cpttt)2][B(C6F5)4], and the relevant vibrational modes are further off-resonance (Figures S24 and S25). Therefore, although chemical alteration of the aromatic rings has made the initial steps in magnetic relaxation slower, confirming our hypothesis,[6a] magnetic relaxation in the Orbach regime in 1 is more efficient than for [Dy(Cpttt)2][B(C6F5)4] due to faster relaxation in the upper energy states of the manifold (Table S10).

Conclusion

In conclusion, we have shown that isolated bis-monophospholyl dysprosium cations can show relatively high Ueff and Tmax values, in common with the cationic bis-cyclopentadienyl dysprosium family. Despite the lack of ring protons in [Dy(Dtp)2][Al{OC(CF3)3}4], and its effective magnetization barrier being identical to that of [Dy(Cpttt)2][B(C6F5)4],[6a] the maximum hysteresis temperature of [Dy(Dtp)2][Al{OC(CF3)3}4] is 12 K lower than this literature example. Ab initio calculations indicate that the replacement of aromatic C5 rings with C4P analogues has slowed down transitions out of the ground |±15/2⟩ doublet as intended. However, smaller energy gaps between excited states that are on-resonance with a series of vibrational modes have rendered relaxation more efficient overall in [Dy(Dtp)2][Al{OC(CF3)3}4]. Therefore, as with the bis-cyclopentadienyl dysprosium cation family,[6,7] the efficacy of magnetic relaxation processes in isolated bis-phospholyl dysprosium cations is also not trivially predictable. This is crucial new information for the future design of lanthanide SMMs with higher magnetic blocking temperatures.

Experimental Section

Materials and Methods

All manipulations were conducted under argon with the strict exclusion of oxygen and water by using Schlenk line and glovebox techniques. Benzene was dried by refluxing over potassium and was stored over a potassium mirror. Chlorobenzene was dried over CaH2 and was stored over 4 Å molecular sieves. All solvents were degassed before use. For NMR spectroscopy, C6D5Cl was dried by refluxing over CaH2 and was vacuum transferred and degassed by three freeze–pump–thaw cycles before use. [Dy(Dtp)2(I)][11d] and Li[Al{OC(CF3)3}4][12] were prepared according to literature methods, and DyI3 (Alfa Aesar) and C3H5MgCl (Sigma-Aldrich) were purchased and were used as received. 1H (400 MHz), 13C (100 and 125 MHz), 31P (162 MHz), and 19F (376 MHz) NMR spectra were obtained on an Avance III 400 or 500 MHz spectrometer at 298 K. These were referenced to the solvent used or to external TMS (1H, 13C), H3PO4 (31P), or C7H5F3/CDCl3 (19F). FTIR spectra were recorded as microcrystalline powders using a Bruker Tensor 27 ATR-Fourier transform infrared (ATR-FTIR) spectrometer. Elemental analysis was carried out by Mr. Martin Jennings and Mrs. Anne Davies at the Microanalytical Service, School of Chemistry, University of Manchester.

[Dy(Dtp)2][Al{OC(CF3)3}4] (1)

A slurry of DyI3 (0.4997 g, 0.92 mmol) and DtpK (0.5311 g, 2.02 mmol) in toluene (20 mL) was heated under reflux for 48 h. The resultant yellow reaction mixture was allowed to cool to room temperature and filtered; the remaining solids were washed with toluene (20 mL). A solution of (C3H5)MgCl in THF (2.0 M, 0.7 mL, 1.4 mmol) was added to the yellow filtrate and stirred for 1.5 h to give an orange reaction mixture. The solvents were removed in vacuo to give a sticky orange solid, which was triturated with a mixture of n-hexane and dioxane (20:1, 30 mL). The product was extracted into n-hexane (15 mL) and filtered, and solvents were removed in vacuo to give an orange foam (0.4100 g, 0.63 mmol, 69% crude yield of the putative “[Dy(Dtp)2(C3H5)]”). [NEt3H][Al{OC(CF3)3}4] (0.6741 g, 0.63 mmol) and benzene (15 mL) were added, and the yellow-orange reaction mixture was stirred overnight. Volatiles from the resultant orange oil and yellow solution were removed in vacuo. The yellow foam obtained was washed with n-hexane (20 mL) and benzene (15 mL), and the residual solvent was removed in vacuo to give a yellow foam. The product was extracted into chlorobenzene (15 mL), filtered, and reduced in volume to 10 mL, and then layered with n-hexane (35 mL). After the reaction mixture was left standing for 3 days at room temperature, large yellow crystalswere obtained; these were washed with n-hexane and dried to give 1 (0.3764 g, 26% global yield based on DyI3). Anal. Calcd (%) for C44H48AlDyF36O4P2: C, 33.51; H, 3.07. Found: C, 30.86; H, 2.78. Elemental analysis results consistently gave lower carbon values than predicted, which we attribute to carbide formation from incomplete combustion. However, all other analytical data obtained are consistent with the bulk purity of 1. χT product = 14.28 cm3 mol–1 K (Evans method). 19F NMR (C6D5Cl): δ = −90.50 (br, v1/2 = 300 Hz). The paramagnetism of 1 precluded assignment of its 1H, 13C, and 31P NMR spectra. FTIR (ATR, microcrystalline; st = strong): ν̃ = 2964 (w, br), 1472 (w), 1397 (w), 1352 (w), 1297 (m), 1274 (m), 1239 (m), 1210 (st), 1165 (m), 1022 (w), 970 (st), 832 (w, br), 726 (st), 660 (w), 624 (w), 560 (w), 536 (m), 442 (m) cm–1.

[NEt3H][Al{OC(CF3)3}4]

A slurry of Li[Al{OC(CF3)3}4] (9.7404 g, 10.0 mmol) and NEt3HCl (1.3765 g, 10.0 mmol) in DCM (175 mL) was stirred overnight. The resultant colorless suspension was filtered, and the solvent was removed from the filtrate in vacuo to give a white powder (7.3038 g, 68%). This was used without further purification; on one occasion the product was recrystallized from a saturated DCM solution and stored overnight at −35 °C, and the solid-state structure was determined by single-crystal XRD (see Supporting Information). Anal. Calcd (%) for C22H16AlF36N: C, 24.71; H, 1.51; N, 1.31. Found: 24.71; H, 1.47; N, 1.46. 1H NMR (CD2Cl2): δ = 1.44 (t, JHH = 7.3 Hz, 9H, NCH2CH3), 3.3 (q, JHH = 7.3 Hz, 6H, NCH2CH3), 4.92 (t, JNH = 54 Hz, 1H, NH). 13C NMR (CD2Cl2): δ = 9.45 (m, NCH2CH3), 49.04 (NCH2CH3), 78.75 (br, s, OC(CF3)2), 121.22 (q, JCF = 293 Hz). 19F NMR (CD2Cl2): δ = −75.70. FTIR (ATR, microcrystalline): ν̃ = 3252 (w), 2986 (w, br), 1746 (w), 1397 (w), 1353 (w), 1271 (s), 1240 (st), 1200 (m), 1024 (w), 967 (st), 833 (w), 797 (w), 756 (w), 725 (st, s), 561 (m), 536 (m), 439 (m) cm–1.