Highly near-IR emissive ytterbium(iii) complexes with unprecedented quantum yields

We report a series of highly NIR emissive Yb complexes, in which the Yb is sandwiched between an octafluorinated porphyrinate antenna ligand and a deuterated Kläui ligand, and one of the complexes has an unprecedented quantum yield of 63%.


Introduction
Highly NIR emissive trivalent lanthanide compounds such as Er, Nd and Yb complexes have been sought aer by generations of chemists because of their many potential applications in telecommunications, 1 light-emitting devices 2 and biological imaging. 2b,3 However the forbidden electric dipole f-f transitions make this task difficult. 4 One way to break this bottleneck is to take advantage of the light-harvesting ability of organic uorophores to sensitize the lanthanide emission via energy transfer, but it has been shown to be quite challenging to implement this strategy. 5 For example, the highest overall quantum yield reported for ytterbium complexes is 12%, achieved with a deuterated ytterbium cryptate in perdeuterated methanol by Seitz et al. 6 Thus it is still of importance to seek an appropriate sensitization system for further improving the NIR emission signicantly.
According to eqn (1), which denes the overall quantum yield F L Ln , an ideal antenna molecular system should meet both the requirements of a high sensitization efficiency (h) and large intrinsic quantum yield (F Ln Ln ). Toward this goal, most reports had focused on increasing either the sensitization efficiency or the intrinsic quantum yield through molecular design. For the sensitization efficiency, porphyrinates (Por) had been shown to be "ideal" candidates among the diverse antenna ligands because of (1) the intense absorption in the visible region; (2) tunable triplet states above the excited state levels of NIR emissive Yb ions with an energy gap in the 2000-3000 cm À1 range; and (3) a good chelating ability. 7 Despite a high sensitization efficiency (close to 100%) being achieved, the overall luminescence quantum yields of Yb porphyrinates are still low (<5%), 8 most likely due to detrimental C-H oscillators in the vicinity of the Yb ion lowering the intrinsic quantum yields. 9 For the intrinsic quantum yield, as early as the 1960s, Kropp and Windsor 10 and Horrocks et al. 11 observed that solvent O-H bonds quench Ln emissions, and later Haas, Stein and Würzberg interpreted this quenching effect in the context of energy gap law. 12 Thus, using heavier atoms such as D and F to replace the H atom of X-H (X ¼ C, N, O) oscillators in the antenna ligand became an alternative approach to enhance the a Beijing National Laboratory for Molecular Sciences, State Key Laboratory of Rare Earth Materials Chemistry and Applications, College of Chemistry and Molecular Engineering, Peking University, Beijing, 100871, P. R. China. E-mail: gaosong@pku. edu.cn; zhangjunlong@pku.edu.cn quantum yields through extending the luminescence lifetime s obs . These early ndings led to the design of many successful ligands for NIR emissive Ln complexes, such as bis(per-uorooctylsulfonyl)amide, 13 bis(pentauorophenyl)phosphinic acid, 14 peruorinated imidodiphosphinate (s obs up to millisecond for Yb), 15 perdeuterated cryptands 6,16 and so on. In addition, shortening the radiative lifetime (s rad ) by changing the coordination sphere could also increase F Ln Ln . 6,17 However, these reported complexes have a low absorption in the visible region or a high-lying energy-donor state, which are disadvantageous for efficient energy transfer from the ligands to the NIR emissive Ln center. Therefore, on the basis of the tremendous progress in the optimization of h and F Ln Ln [eqn (1)], integrating the best of these components by using a good light harvesting antenna such as a porphyrin and depleting the X-H oscillators is consequently a promising approach for achieving highly NIR emissive lanthanide complexes.
In this work, we report that a high sensitization efficiency and a large intrinsic quantum yield can be simultaneously obtained using a molecular system with a peruorinated porphyrin, 2, 3,7,8,12,13,17,18-octauoro-5,10,15,20-tetrakis(pentauorophenyl)porphyrin (1), as the sensitizer and a partially deuterated Kläui's tripodal ligand [D 18 ]-L OMe , [(cyclopentadienyl)tris(di(methyl-d 3 )phosphito)cobaltate] 18 as the capping ligand. Using this approach, we are able to dramatically increase the overall quantum yield of the corresponding Yb(III) complex [D 18 ]-1-Yb up to 63% (estimated uncertainty 15%) and extend the lifetime to 714 ms in CD 2 Cl 2 . Given the large extinction coefficient of the porphyrinate (3 $ 320 000 M À1 cm À1 ) in the visible to red region, [D 18 ]-1-Yb represents one of the brightest NIR-emissive Ln(III) complexes (brightness: 3 Â F L Ln $ 190 000 M À1 cm À1 ) ever reported upon excitation in the visible range (>400 nm). Systematic analysis of the structure-photophysical properties relationship suggests the importance of both a porphyrinate antenna ligand and a C-H oscillator free coordination environment for designing highly NIR emissive Yb complexes. Furthermore, the subtle effects that meso-phenyl groups of the b-octauorinated porphyrins have on the NIR emission properties of Yb(III) complexes are anticipated to enable synthetic exibility and practicality for the further design of functional materials that emit in the NIR region.
Three representative ytterbium complexes 1-Yb, 2-Yb and [D 18 ]-3-Yb were also crystallographically characterized ( Fig. 1B-D). In these complexes, the Yb ion is seven-coordinate, and surrounded by four N atoms from the porphyrinate dianion and three O atoms from the Kläui ligand. The three mean planes (C5 of the cyclopentadienyl ring, N4 of the porphyrin, and O3 of the phosphito groups) are almost parallel to each other. The bond lengths of Yb-N (2.35-2.38Å) and Yb-O (2.20-2.27Å) and the distances between the Yb and the N4 mean plane (1.16-1.19Å) in the three compounds are very similar, with a difference of < 0.1Å (ESI Table S2 †), consistent with similar previously reported ytterbium porphyrinate complexes. 8a,19a,21 Photophysical properties The photophysical properties of the Yb complexes were investigated using UV-vis absorption and photoluminescence spectroscopy. All the complexes displayed an intense Soret band at 380-450 nm and moderately intense Q bands at 520-600 nm in dichloromethane, which are the characteristic absorptions of Ln(III) porphyrinates ( Fig. 2  Excitation within the Soret band region at 425 nm yields a typical NIR emission centred at 974 nm for all the Yb(III) complexes, which is assigned to the Yb( 2 F 5/2 / 2 F 7/2 ) transition. The excitation spectra are consistent with the corresponding absorption spectra, suggesting energy transfer from the porphyrinate ligands to the Yb(III) centers (ESI, Fig. S58-67 †). Emission in the visible region (520-780 nm) was also observed, which was much weaker compared to the free bases, again pointing to the energy absorbed by the porphyrinate chromophore being efficiently transferred to the metal centers. 8a,d,22a,23 Interestingly, under the same photoluminescence conditions, the emission intensity was found to decrease in the order 18 ]-2-Yb > 2-Yb, indicating that b-uorination of the porphyrin and deuteration of the Kläui ligand improve the NIR emission of Yb signicantly (Fig. 3A). Accordingly, the lifetimes of these Yb complexes follow the same order as the luminescence intensity (Fig. 3B).
In order to quantify these results, the NIR emission quantum yields (F L Yb ) and lifetimes (s obs ) of the Yb(III) complexes were measured and the data are summarized in Table 1. It is worth noting that the estimated uncertainties in F L Yb are 15%. 24 The F L Yb values for the NIR emission were obtained using a comparative method with Yb(TPP)(L OEt ) (F ¼ 2.4% in CH 2 Cl 2 , H 2 TPP ¼ 5,10,15,20-tetraphenylporphyrin; 8a as a reference, using a FLS920 spectrometer (Edinburgh Instruments) equipped with a PMT R5509-73 detector (300-1700 nm, Hamamastu) for NIR emission. Among all the Yb complexes, [D 18 ]-1-Yb displays overall quantum yields of 25% in CH 2 Cl 2 and 63% in CD 2 Cl 2 (ESI, Fig. S87 †), which are much higher than the quantum yield of the Yb(III) cryptate reported by Seitz et al. in 2015 (12% in CD 3 OD). 6 These values were further conrmed using a real photon counting method, using an integrating sphere and the same instrument, which gave F L Yb values of 26% in CH 2 Cl 2 and 69% in CD 2 Cl 2 respectively (ESI, Fig. S88-S89 †). Since the sensitivity of the PMT detector is lower within the optical window of 900-1100 nm than a CCD detector, 25 we also measured the quantum  yields of [D 18 ]-1-Yb using a Fluorolog-3 spectrouorimeter equipped with a CCD detector (Horiba Scientic, see Experimental section for details). 25 In CH 2 Cl 2 , the quantum yield of [D 18 ]-1-Yb was 23%, close to that obtained using the PMT 5509-73 detector. The quantum yield in CD 2 Cl 2 was 49%, which is lower than that obtained using the PMT detector, probably due to the CCD detector having a cutoff at 1100 nm, whereas [D 18 ]-1-Yb still has a strong emission beyond this region (Fig. 3A). Nevertheless, aer consideration of the quantum yields obtained using different methods, the overall quantum yields of [D 18 ]-1-Yb in CH 2 Cl 2 and CD 2 Cl 2 are still much higher than those of the previously reported Yb complexes. 26 The lifetimes of [D 18 ]-1-Yb are as long as 285 ms in CH 2 Cl 2 and 714 ms in CD 2 Cl 2 , and much longer than those of most reported ytterbium complexes (typically below 100 ms). 8c,15b,27 Taking the high extinction coefficient of the porphyrinate ligand (3 414 nm z 320 000 M À1 cm À1 ) into account, [D 18 ]-1-Yb represents one of the brightest NIR emissive Ln complexes (brightness: 3 Â F L Yb $ 190 000 M À1 cm À1 ) ever reported upon excitation in the visible range (>400 nm).
The   ]-2-Yb (>2 fold increase of F L Ln on going from CH 2 Cl 2 to CD 2 Cl 2 ), compared to a much smaller effect observed for 1-Yb, 2-Yb and [D 18 ]-2-Yb (+4 to +31%). We conclude that, in the absence of high-energy C-H oscillators in the inner coordination sphere, the emissive state of Yb(III) is deactivated mainly through solvent C-H vibrational relaxation, as substantiated by the signicant isotopic solvent effect observed.
To better elucidate the deactivation ability of C-H oscillators, we have elaborated a quenching sphere model for use with the [(Por)Yb(L OR )] complexes studied (Fig. 4). The estimated quenching rate differences Dk (in ms À1 ) for C-(H/D) or C-(H/F) at different sites, calculated from the lifetimes of the corresponding complexes, are also shown (see ESI for details, Table S5 †). According to the Dk values, the coordination sphere is divided into three layers, within which the gradually fading blue color indicates weakening of the quenching effects of the C-H oscillators in the sphere (not strictly distance-dependent). The coordinating N and O atoms are in the primary sphere. The C-H oscillators at b-pyrrolic positions of the porphyrinate (Dk F/H ¼ 13.5 ms À1 and Dk D/H ¼ 12.8 ms À1 ) and the phosphito groups of the Kläui ligands (Dk ¼ 16.1 ms À1 ) with a large Dk value t into the second sphere. They account for the major vibrational quenching contribution in the Yb complexes. The C-H oscillators of the meso-phenyl groups play a role in the third sphere, with much smaller Dk values (0.8 ms À1 for ortho-positions and 0.4 ms À1 for meta-and para-positions respectively). The higher quenching rate for ortho-C-H bonds than that of meta-or para-C-H bonds is probably due to the shorter distances to the Yb center for the former (5.00-6.21Å) compared to those of the latter (7.33-8.86Å). The C-H bonds of solvent molecules interact with the excited state of Yb formally in the third sphere with Dk D/H (Dk D/H ¼ 2.6 ms À1 ) comparable to that of the mesophenyl ones.

Sensitization efficiency and intrinsic quantum yield
Having established the inuence of C-H oscillators on the emission properties of the ytterbium porphyrinate complexes, we started to investigate the key factors governing improvement of the luminescence efficiency for the present molecular system. According to eqn (1), the overall quantum yield F L Yb is determined by two components: the sensitization efficiency (h) and intrinsic quantum yield F Yb Yb . The latter is calculated from the ratio of non-radiative deactivation process (s obs ) to the radiative lifetime (s rad ). 26 s rad of Yb(III) can be estimated from f-f transition absorption spectra based on a modied Einstein equation: 32 where c is the speed of light in cm s À1 , n is the refractive index (n(CD 2 Cl 2 ) ¼ 1.442), N A is Avogadro's number, J and J 0 are the  quantum numbers for the ground and excited states, respectively, Ð 3ðỹÞdðỹÞ is the integrated spectrum of the 2 F 7/2 / 2 F 5/2 transition, andỹ m is the barycenter of the transition.
Given that these Yb(III) complexes share identical N 4 O 3 coordination environments, including similar Yb-N and Yb-O bond lengths, the radiative lifetimes of these complexes are not expected to vary greatly. 6,17a Thus we chose [D 18 ]-3-Yb as an example to measure s rad in CD 2 Cl 2 because of its good solubility up to 10 À2 M (ESI Fig. S92 †). The derived radiative lifetime of s rad ¼ 0.95 ms ($15% error), which is within the typical range of 0.5-1.3 ms for ytterbium complexes in solution, 9d,27c,32-33 was used for all the Yb(III) complexes to calculate the sensitization efficiency and intrinsic quantum yield, and the results are tabulated in Table 2.
All the Yb(III) complexes have high h values (>70%) in CD 2 Cl 2 , suggesting that b-uorination or deuteration of the ligands has a subtle effect on the energy transfer process from the lowest triplet state (T 1 ) of the ligands to the Yb(III) excited state. The T 1 energy levels of these compounds were determined from low-temperature emission spectra of the corresponding Gd(III) complexes and are in the range 12 100-13 700 cm À1 , which is considered to be optimum for efficient Yb sensitization. 4b Besides, the short distances between the Yb and the porphyrin N4 mean plane determined from the crystal structure are also favourable for efficient energy transfer. Thus, the sensitization efficiency is not the main factor responsible for the pronounced differences in the NIR emissions found in this work between these Yb complexes. In contrast, the intrinsic quantum yields for the Yb(III) complexes are signicantly different from each other (3.5-75% in CD 2 Cl 2 , estimated error 15%), as a result of disparity in s obs , which is highly related to the degree of C-H oscillator substitution. Among them, [D 18 ]-1-Yb has the highest intrinsic quantum yield ever reported, ca. 75% in CD 2 Cl 2 , owing to its extremely long lifetime. Moreover, plotting F L Yb vs. F Yb Yb shows an approximately linear relationship (Fig. 5), suggesting a decisive role of F Yb Yb in determining the F L Yb of the Yb(III) porphyrinates. Therefore, minimization of the non-radiative processes via uorination and deuteration is the main origin of the increased quantum yields, which reach a maximum for the nearly C-H bond free compound [D 18 ]-1-Yb.

Conclusion
In summary, we report here a molecular system for achieving highly luminescent Yb(III) complexes with a new benchmark quantum yield of 63% (estimated uncertainty 15%). Systematic analysis of the photophysical properties and the structures of the complexes revealed that a C-H bond depleted coordination sphere is critical for obtaining a high NIR emission efficiency, as a result of minimized non-radiative processes. The b-pyrrolic C-H bonds of the porphyrin and the phosphito C-H bonds of the Kläui ligand greatly inuence the Yb(III) luminescence, whereas those of the meso-phenyl group substituents on the porphyrin only have a slight effect. Fluorination of the porphyrin ligand was shown to have a much more benecial effect than deuteration. In addition to the high quantum yield, other attractive features of these compounds such as excitation in the visible range, large extinction coefficients and synthetic exibility make them easily adaptable for the design of potential light converting systems.

Experimental section
General materials and methods UV-vis spectra were recorded using an Agilent 8453 UV-vis spectrometer equipped with an Agilent 89090A thermostat (AE0.1 C) at 25 C. Near-IR absorption spectra were recorded using a Shimadzu UV-3600 Plus UV-Vis-NIR Spectrophotometer. Mass spectra were recorded using a Bruker APEX IV FT-ICR mass spectrometer (ESI-MS). Elemental analyses (C, H, N) were performed using an Elementar Analysensysteme GmbH vario EL Elemental Analyzer. NMR spectra were recorded using a Varian Mercury Plus 300 MHz spectrophotometer or Bruker ARX400 400 MHz spectrophotometer. IR spectra were recorded using a Bruker VECTOR22 FTIR spectrometer and KBr pellets. For the optical measurements in liquid solution, spectroscopic grade CD 2 Cl 2 was purchased from Cambridge Isotope Laboratories, Inc. and used as received. Anhydrous CH 2 Cl 2 was distilled from calcium hydride and 1,2,4-trichlorobenzene (TCB) was purchased from J&K Scientic. The b-octauorinated porphyrin ligands 34 and deuterated Kläui's ligand 18b (D atom > 99%) were synthesized according to literature methods.

Synthesis of lanthanide porphyrinates
The syntheses were carried out according to modied literature methods. 19a,20a Generally, a porphyrin (0.03 mmol) and Ln(acac) 3 $nH 2 O (0.15 mmol) were reuxed in 8 mL of TCB for 2 h under N 2 . During the reaction process, the luminescence of the porphyrin free base gradually vanished. Aer cooling to room temperature, the reaction mixture was eluted with petroleum ether, CH 2 Cl 2 , and CH 2 Cl 2 /MeOH ¼ 5/1 sequentially to provide TCB, the unreacted porphyrin free base and the lanthanide porphyrin complexes, in order, using ash silica gel chromatography. The lanthanide complex and 1.2 equiv. of the Kläui ligand L OR (or the partially deuterated one) were stirred in 10 mL of CHCl 3 /MeOH (1/1) at 60 C for 2 h. Then the product, with the general formula [Ln(Por)(L OR )], was isolated using silica gel chromatography and recrystallized from CH 2 Cl 2 /n-hexane.

Photophysical properties measurement
The emission spectrum and lifetime were recorded using an Edinburgh Analytical Instrument FLS920 lifetime and steady state spectrometer (450 W Xe lamp/microsecond ash lamp, PMT R928 for the visible emission spectrum, PMT R5509-73 with a C9940-02 Hamamatsu cooler for the NIR emission spectrum and luminescence lifetime). All the emission spectra in the NIR region were corrected using a calibration curve for the detector response (Fig. S55 †). The NIR quantum yields of all the complexes were measured using a comparative method with Yb(TPP)(L OEt ) as the reference aer excitation at l ex ¼ 425 nm (2.4%, CH 2 Cl 2 solution). Sample quantum yields were evaluated using the following equation: where the subscripts r and s denote the reference and sample respectively, F is the quantum yield, k is the slope from the plot of integrated emission intensity vs. absorbance, and n is the refractive index of the solvent. The estimated error for the quantum yield measurements is 15%.
The quantum yield of [D 18 ]-1-Yb was also determined using integrating spheres and two instruments. The rst was an integrating sphere (150 mm, PTFE inner surface) tted within the Edinburgh Analytical Instrument FLS920 with a PMT R5509-73 detector for NIR emission and a PMT R928 for visible emission. The second was a Quanta-4 integrating sphere (150 mm, PTFE inner sphere, Horiba Scientic) along with a Horiba-Jobin-Yvon Fluorolog-3 spectrouorimeter equipped with a CCD detector (1024 Â 256 pixel, 200-1100 nm, Horiba Scientic) referenced to Yb(TPP)(L OEt ). The quantum yields determined with the FLS920 were evaluated according to the following equation: where A em is the integrated area of the sample's emission (corrected); A ref scatter and A sample scatter are the integrated areas under the Rayleigh scattering peaks of the reference sample and the sample under study; and k R928/NIR is the ratio of the sensitivities of the two detectors. The value of k R928/NIR was determined straight aer the measurement.