doi.org/10.26434/chemrxiv.7067669.v1

Correlating Blocking Temperatures in Single Molecule Magnets with Raman Relaxation

Marcus J. Giansiracusa, Andreas Kostopoulos, George F. S. Whitehead, David Collison, Floriana Tuna, Richard Winpenny, Nicholas Chilton

Submitted date: 10/09/2018 • Posted date: 11/09/2018 Licence: CC BY-NC-ND 4.0 Citation information: Giansiracusa, Marcus J.; Kostopoulos, Andreas; F. S. Whitehead, George; Collison, David; Tuna, Floriana; Winpenny, Richard; et al. (2018): Correlating Blocking Temperatures in Single Molecule Magnets with Raman Relaxation. ChemRxiv. Preprint.

We report a six coordinate DyIII single-molecule magnet (SMM) with an energy barrier of 1110 K for thermal relaxation of magnetization. The sample shows no retention of magnetization even at 2 K and this led us to find a good correlation between the blocking temperature and the Raman relaxation regime for SMMs. The key parameter is the relaxation time (ᵰ ) at the point where switch the Raman relaxation mechanism becomes more important than Orbach.

File list (1)

Dy-mon_final_for ChemRxiv.pdf (1.77 MiB) view on ChemRxiv download file

Correlating Blocking Temperatures in Single Molecule Magnets with Raman Relaxation Marcus J. Giansiracusa, Susan Al-Badran, Andreas K. Kostopoulos, George F. S. Whitehead, David Collison, Floriana Tuna, Richard E. P. Winpenny*, and Nicholas F. Chilton*

Dedications

Abstract: We report a six coordinate DyIII single-molecule magnet length to the anionic DiMeQ oxygen donors at 2.150(4) Å. The (SMM) with an energy barrier of 1110 K for thermal relaxation of trans equatorial Dy-Cl bonds are 2.681(2) Å, and the third Cl magnetization. The sample shows no retention of magnetization even ligand trans to the neutral water ligand (2.32(1) Å) has a bond at 2 K and this led us to find a good correlation between the blocking length of 2.897(8) Å. There are strong intramolecular N-H…Cl temperature and the Raman relaxation regime for SMMs. The key hydrogen bonds of 2.322(2) Å which support the near perfect parameter is the relaxation time ( � ) at the point where the octahedral coordination geometry (validated using Shape Raman relaxation mechanism becomes more important than Orbach. software, see ESI).[8,9] There are intermolecular hydrogen-bonding interactions Several lanthanide-based single molecule magnets (SMMs) between the water and a bound chloride on the adjacent have now been reported with energy barriers for reversal of [1–5] molecule (O-H…Cl 2.244(9) Å) leading to chains running through magnetization (Ueff) greater than 1000 K. However, the the structure; the Dy…Dy distance is 7.1829(6) Å along these temperature at which these high-barrier SMMs retain chains. p-stacking of the DiMeQ ligands interlocks these chains, magnetization differ markedly. The dysprosocenium cation ttt ttt t where the closest C…C contacts between carbon atoms of [Dy(Cp )2][B(C6F5)4] (Cp = C5H2 Bu3-1,2,4), shows magnetic adjacent chains is 3.382(9) Å (Figure 1b). The strong hysteresis up to 60 K and has a blocking temperature (TB) of 40 [1] t intermolecular interactions lead to a very rigid, closely packed K, while [Dy(O Bu)2(py)5][BPh4] has a higher Ueff barrier yet [4] 2D structure. exhibits hysteresis only up to 4 K, and TB = 14 K. Here we use the conventional definition of TB as the peak in the zero-field cooled (ZFC) magnetic susceptibility of a superparamagnet.[6] One feature of these new high-barrier SMMs is that even this simple definition of TB has become questionable as the observed [7] behavior near TB varies significantly.

Here we report a compound with a very high Ueff that shows negligible hysteresis even at 2 K. The compound has the formula

[Dy(DiMeQ)2Cl3(H2O)] (DiMeQ = 5,7-dimethyl-8- hydroxyquinoline, Figure 1) 1, and was prepared through reaction of DyCl3.6H2O and 5,7-dimethyl-8-hydroxyquinoline in a molar ratio of 1:2 in methanol (see ESI for details); the isostructural [Y(DiMeQ)2Cl3(H2O)] 1Y, can be prepared by a similar route. Compound 1 crystallizes in P-1 and features a six-coordinate DyIII ion bound to two trans DiMeQ ligands through the deprotonated phenoxide group, three mer chloride anions and a water molecule (Figure 1a). The coordinating DiMeQ ligand has a neutral charge due to the deprotonation of the phenoxide and the protonation of nitrogen in the quinolate ring. The centrosymmetric structure with trans-disposition of the DiMeQ ligands gives an O-Dy-O bond angle of 180 ° with the Dy-O bond Figure 1. (a) Crystal structure of 1 with the intra-molecular H-bonds. (b) Packing of 1 showing the inter-molecular H-bonds and p-stacking interactions. H-atoms not involved in H-bonding omitted for clarity [a] M. J. Giansiracusa, S. Al-Badran, Dr. A. K. Kostopoulos, Dr. G. F. S. Whitehead, Prof. D. Collison, Dr. F. Tuna, Prof. E.J. L. McInnes, 3 -1 Prof. Richard E. P. Winpenny*, and Dr. Nicholas F. Chilton Magnetic studies of 1 give a χMT product of 13.2 cm mol K at The School of Chemistry, The University of Manchester 300 K, as expected for a DyIII ion (Figure S3).[10] The Oxford Road, Manchester, M13 9PL, United Kingdom magnetization data saturate to a value of 4.9 N μ by 3 T, Email: [email protected], A B [email protected] consistent with an mJ = ±15/2 ground state. AC susceptibility measurements in zero field reveal strong frequency-dependent S. Al-Badran peaks in the out-of-phase susceptibility up to 68 K (Figure S4). Chemistry Department, College of Science, Basrah University, Basrah, Iraq Fitting the Cole-Cole data from 2 – 68 K to a generalized Debye model yields low alpha values (α < 0.2) indicating a single Supporting information for this article is given via a link at the end of the document. relaxation process (Figure S4-5, Table S3). Fitting the relaxation

-12 rates to Equation 1 yields Ueff = 1110(50) K with τ0 = 2(1) x 10 unobservable due to the weak signal at high temperatures -3 -1 -n s, C = 5(1) x 10 s K , n = 3.32(7) and τQTM = 0.0244(9) s (Figure 2). Furthermore, we observe a shift in the quantum

(Figure 2, S6). tunneling of magnetization (QTM) to a slower rate of τQTM = 2.0(3) s, indicating that dipolar fields have an influence on QTM (Figure 2, S9-10, Table S6). Hysteresis measurements reveal a slight opening of the loops (Figure S11) along with a maximum in the = � + �� + (1) ZFC curve at 6 K (Figure S12). The environment of the Dy site in 1 is highly anisotropic and stabilizes the large |mJ| projections of the ground Dy(III) multiplet, as shown by an electrostatic calculation (Figure S7).[11,12] To obtain quantitative insight into the electronic structure, we performed complete active space self-consistent field spin-orbit (CASSCF-SO) calculations on the structure of 1. The low lying crystal field (CF) states are almost pure mJ functions (Table S4); consequently, the first two doublets have highly axial principal g- values with minimal deviation in the directions of the largest g- value for the two Kramers doublets. Magnetic relaxation is most likely to occur via the third excited state which has an almost easy-plane g-tensor; this state has an energy of 1037 K (726 cm- 1) above the ground state, in excellent agreement with the experimental energy barrier obtained from fitting the AC data. ZFC/FC measurements with an applied field of 100 Oe yield curves that almost perfectly overlay (Figure S8), and hysteresis Figure 3. Magnetic hysteresis measurements of 1 performed at 1.8 and 5 K showing no remanent magnetization at either temperature, insets with zoom measurements show rapid closing of hysteresis loops at zero- around 0 T for clarity. Field sweep rate of ~15 Oe/s. field with no remanent magnetization (Figure 3). These data indicate that there is no magnetic blocking observable at 1.8 K Despite dilution, the performance of this SMM is unexpectedly 1 or higher temperatures; thus TB < 1.8 K. poor given Ueff > 1000 K. We examined if the nearby H nuclei of the bound water molecule had any specific influence on the SMM properties by preparing a partially deuterated sample using

CD3OD as a solvent with DyCl3·6D2O in the synthetic process to

obtain 1d. The presence of the D2O ligand was confirmed by IR spectroscopy (Figure S1), revealing the absence of the O-H stretch. However, magnetic measurements of 1d are nearly identical to 1 (Figure S13-18, Table S7), thus hyperfine coupling to these nearby 1H nuclei is not the source of the poor performance. We are also confident that 161/163Dy hyperfine is not responsible based on our recent work removing the Dy-based nuclear spins in SMMs.[13] While 1 joins the few reported SMMs with energy barrier over 1000 K (Table 1, S7 and Figure S19), it does not retain magnetization: we felt understanding why 1 is so poor might help develop better SMMs. There is clearly no good correlation Figure 2. Combined fitting of relaxation rate data for 1 (black points) and 1@1Y (red points) using the same Orbach and Raman parameters with unique QTM. between Ueff and TB (Figure S20); the short-coming of Ueff as the Orbach regime (black), Raman (red), QTM pure (light blue) and dilute (dark blue) defining figure of merit has been highlighted previously in the and overall fits for pure (light green) and dilute (dark green). literature for lower barrier systems.[14,15]. It has been postulated that a flexible lattice is responsible for rapid relaxation observed in poor performing SMM systems;[16] however, due to the rigid III We probed the effect of dipolar interactions between Dy sites network of hydrogen bonds and p-stacking interactions present through dilution experiments. A doped sample (1@1Y) was in 1, this cannot be the explanation here. synthesized with a 1:19 molar ratio between DyCl3.6H2O and In an attempt to understand why 1 is such a bad SMM, we have YCl3.6H2O, and purity of the bulk sample was confirmed by examined the Raman relaxation regime for high Ueff SMMs and powder x-ray diffraction (Figure S2); elemental analysis shows this shows a trend. Taking the reported Raman parameters for around 7% concentration of Dy. This concentration results in III diamagnetically diluted monometallic Dy SMMs with Ueff >≈ partial isolation of the sites, with around 56% expected to have 1000 K (to minimize the number of different factors to which no paramagnetic neighbors (binomial distribution across the 8 relaxation dynamics may be attributed, Table 1), we have nearest-neighbors). The AC data for 1@1Y yield an almost calculated and plotted relaxation times (τ) between 15 and 40 K identical Raman region to 1, with Raman parameters of on a log-log scale (Figure 4); note that for all compounds the C = 5(1) x 10-4 s-1 K-n and n = 3.94(7), but the Orbach process is

Raman mechanism is dominant in the 15 – 40 K region and for equal (i.e. the switching point). Indeed, a plot of log[TB] vs. t [Dy(O Bu)2(py)5][BPh4] that the parameters for the pure analogue log[τSwitch] shows a linear correlation with log � = [4] 2 have been used as no dilution experiments were performed. 0.15(2)log � + 1.25(6) with R = 0.88 (Figure 5), We have omitted bimetallic systems with significant magnetic suggesting that the magnetic blocking in DyIII SMMs has far more [3,15] interactions, however we include [Dy2(Cp*)2(FeCp(CO)2)2] to do with Raman relaxation as the Ueff value is now so high the due to the significant isolation of the Dy ions as a result of the Orbach mechanism is irrelevant (Figure S20). [17] long {FeCp2(CO)2} bridges. In comparing the Raman relaxation rates with TB, we define TB using the peak in ZFC measurements, except for [Dy2(Cp*)2{FeCp(CO)2}2] where

ZFC/FC was not measured; thus, it was assigned TB = 6 K by hysteresis measurements with a sweep rate of 20 Oe/s.[17]

Table 1. A list of SMMs from literature with energy barriers close to and over 1000 K with parameters obtained from dilution experiments, except for t [Dy(O Bu)2(py)5][BPh4] where dilution was not performed.

Ueff TB ������� Eq. REF (K) (K) (s) ligands

ttt [Dy(Cp )2] 1.1 1754 40 2 None 1 [B(C6F5)4] x10

t Figure 5. Relaxation times calculated at the intersection point of the Obarch and [Dy(O Bu)2(py)5] 9.9 Raman fitting parameters reported for high barrier SMMs plotted against 1815 14 -2 5 N 4 ttt + [BPh4] x10 literature TB values on a log-log plot. A linear relation is evident with [DyCp 2] at the top, then other systems falling into order of blocking temperature. The 2.0 open circle is relaxation rate calculated for the lower barrier system [Dy(bbpen)Br] 1191 8 -3 4N, Br 2 t x10 [ BuPO(NHiPr)2Dy(H2O)5].

t [Dy(O Bu)Cl(THF)5] 1.7 950 7 -2 5 THF 5 The next step would be to link the structures of these SMMs [BPh4] x10 to their Raman relaxation rates, which in turn would imply a [Dy (Cp*) 4.4 2 2 correlation with TB. Immediately obvious is the unique structure 953 6 -3 2(OC) 17 {FeCp(CO)2}2] x10 of the system with highest TB: the coordination environment of [Dy(Cpttt)2][B(C F ) ] is approximately axial with no equatorial [Dy(DiMeQ)2Cl3 1.9 H2O, This 6 5 4 1110 6 -4 ligands; the relaxation dynamics have been thoroughly (H2O)] x10 3Cl work investigated, showing the importance of the restricted vibrational a. Cpttt = {C H tBu -1,2,4} b. H bbpen = N,N‘-bis(2- hydroxybenzyl)-N,N‘-bis(2- 5 2 3 2 modes of the bis-Cpttt coordination geometry on the Raman methylpyridyl)ethylenediamine c. Cp* = pentamethylcyclopentadienyl parameters.[1,18,19] Figure 4 clearly shows a correlation between TB and the Raman The two next best examples (Table 1) contain predominantly relaxation time in the range of 15 – 40 K, with the highest TB N-donors in the equatorial plane [Dy(OtBu) (py) ][BPh ] and SMMs having significantly longer relaxation times. While a trend 2 5 4 [Dy(bbpen)Br], however the latter also has an equatorial bromide is evident, any numerical correlation would involve choosing a ion. The next two on the list contain O-donors, including the specific temperature arbitrarily. As a non-arbitrary metric with [17] unusual O-bound carbonyl in [Dy2(Cp*)2(FeCp(CO)2)2]. Compound 1 features three equatorial chloride anions. The t nearest to a direct comparison is between [Dy(O Bu)2(py)5][BPh4] t and [Dy(O Bu)Cl(THF)5][BPh4], both with five equatorial

monodentate ligands, and here TB for the N-donor is twice that for the O-donor. There is therefore some correlation with the electronegativity of the donor atom of the equatorial ligands, however further work will be needed to establish this definitively.

To see if our obtained correlation is consistent with lower Ueff barriers, we examined a system with a lower barrier where Raman parameters have been reported. The compound t -4 -1 -n Figure 4. Relaxation times from the Raman fitting parameters reported for high [ BuPO(NHiPr)2Dy(H2O)5] has Ueff = 651 K, C = 6.92 x 10 s K barrier SMMs plotted between 15 and 40 K. A clear trend is visible with [20] ttt + and n = 3, with a noteworthy TB = 12 K; initially this TB appears [DyCp 2] at the top, then other systems falling almost perfectly into order of - t inconsistent with the lower Ueff barrier. Plotting its τSwitch (6.5 x 10 blocking temperature, with the only exception being [Dy(O Bu)Cl(THF)5][BPh4] 2 lying above [Dy(bbpen)Br]. s) and TB values onto Figure 5 results in the open circle, which is in excellent agreement with our correlation. which to compare these systems, we propose the point at which This examination of literature SMMs shows a strong the Raman relaxation mechanism takes over from the Orbach correlation between the characteristic relaxation time where regime. Using fitted parameters from literature, we identify the Raman relaxation takes over from the Orbach regime and the relaxation time (�) at which these two competing rates are

magnetic blocking temperature, which seem to be related to the and N. F. Chilton, 2018, DOI: 10.26434/chemrxiv.6790568.v1. number and type of equatorial ligands. This latter magneto- [14] K. S. Pedersen, J. Dreiser, H. Weihe, R. Sibille, H. V Johannesen, structural correlation needs further study of SMMs in strongly M. a Sørensen, B. E. Nielsen, M. Sigrist, H. Mutka, S. Rols, J. axial crystal fields, with systematic variation in the ligands bound Bendix, S. Piligkos, Inorg. Chem. 2015, 54, 7600–7606. in the equatorial sites. [15] E. Rousset, M. Piccardo, M.-E. Boulon, R. Gable, A. Soncini, L. Sorace, C. Boskovic, Chem. - A Eur. J. 2018, DOI Acknowledgements 10.1002/chem.201802779. [16] A. Lunghi, F. Totti, R. Sessoli, S. Sanvito, Nat. Commun. 2017, 8, This work was supported by The University of Manchester, and 14620. the Ramsay Memorial Fellowships Trust (fellowship to NFC). We [17] T. Pugh, N. F. Chilton, R. A. Layfield, Angew. Chemie Int. Ed. 2016, thank The University of Manchester for access to the SQUID 55, 11082–11085. magnetometer in the EPSRC National EPR Facility, and the [18] C. A. P. Goodwin, D. Reta, F. Ortu, N. F. Chilton, D. P. Mills, J. Am. EPSRC for funding an X-ray diffractometer (grant number Chem. Soc. 2017, 139, 18714–18724. EP/K039547/1). M.J.G. thanks The University of Manchester for [19] C. A. P. Goodwin, D. Reta, F. Ortu, J. Liu, N. F. Chilton, D. P. Mills, a President’s Doctoral Scholarship. S. A.-B. thanks the Higher Chem. Commun. 2018, 54, 9182–9185. [20] S. K. Gupta, T. Rajeshkumar, G. Rajaraman, R. Murugavel, Chem. Committee for Education Development in Iraq (HCED) for the Sci. 2016, 7, 5181–5191. award of a research scholarship.

Conflict of interest

The authors declare no conflict of interest.

Keywords: Single-molecule magnets, dysprosium, blocking temperature, energy barrier, Raman

[1] C. A. P. Goodwin, F. Ortu, D. Reta, N. F. Chilton, D. P. Mills, Nature 2017, 548, 439–442. [2] J. Liu, Y.-C. Chen, J.-L. Liu, V. Vieru, L. Ungur, J.-H. Jia, L. F. Chibotaru, Y. Lan, W. Wernsdorfer, S. Gao, X.-M. Chen, M.-L. Tong, J. Am. Chem. Soc. 2016, 138, 5441–5450. [3] D. S. Krylov, F. Liu, S. M. Avdoshenko, L. Spree, B. Weise, A. Waske, A. U. B. Wolter, B. Büchner, A. A. Popov, Chem. Commun. 2017, 53, 7901–7904. [4] Y.-S. Ding, N. F. Chilton, R. E. P. Winpenny, Y.-Z. Zheng, Angew. Chemie Int. Ed. 2016, 55, 16071–16074. [5] Y.-S. Ding, K.-X. Yu, D. Reta, F. Ortu, R. E. P. Winpenny, Y.-Z. Zheng, N. F. Chilton, Nat. Commun. 2018, 9, 3134. [6] D. Gatteschi, R. Sessoli, J. Villain, Molecular Nanomagnets, Oxford University Press, Oxford, 2006. [7] J.-L. Liu, Y.-C. Chen, M.-L. Tong, Chem. Soc. Rev. 2018, 47, 2431– 2453. [8] D. Casanova, J. Cirera, M. Llunell, P. Alemany, D. Avnir, S. Alvarez, J. Am. Chem. Soc. 2004, 126, 1755–1763. [9] P. . Llunell, M.; Casanova, D.; Cirera, J.; Bofill, J. M.; Alemany, D. Alvarez, S.; Pinsky, M.; Avnir, SHAPE 2.0, Universitat De Barcelona And The Hebrew University Of Jerusalem, Barcelona and Jerusalem, 2003. [10] C. Benelli, D. Gatteschi, Introduction to Molecular Magnetism, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2015. [11] N. F. Chilton, D. Collison, E. J. L. McInnes, R. E. P. Winpenny, A. Soncini, Nat. Commun. 2013, 4, 2551. [12] M. J. Giansiracusa, E. Moreno-Pineda, R. Hussain, R. Marx, M. Martínez Prada, P. Neugebauer, S. Al-Badran, D. Collison, F. Tuna, J. Van Slageren, S. Carretta, T. Guidi, E. J. L. McInnes, R. E. P. Winpenny, N. F. Chilton, J. Am. Chem. Soc. 2018, 140, 2504–2513. [13] F. Ortu, D. Reta, Y.-S. Ding, C. A. P. Goodwin, M. P. Gregson, E. J. L. McInnes, R. E. P. Winpenny, Y.-Z. Zheng, S. T. Liddle, D. P. Mills

Table of Contents

COMMUNICATION

Marcus J. Giansiracusa, Susan Al- Badran, Andreas K. Kostopoulos, George F. S. Whitehead, David Collison, Floriana Tuna, Eric J. L. McInnes, Richard E. P. Winpenny*, and Nicholas F. Chilton*

Page No. – Page No.

Correlating Blocking Temperatures in Single Molecule Magnets with Raman Relaxation

Dy-mon_final_for ChemRxiv.pdf (1.77 MiB) view on ChemRxiv download file