Controlled Under Pressure

DOI: 10.1002/anie.200((will be filled in by the editorial staff))

Pressure-Driven Orbital Reorientations and Coordination Sphere Reconstructions in [CuF2(H2O)2(pyz)]

Alessandro Prescimone, Chelsey Morien, David Allan, John A. Schlueter, Stan W. Tozer, Jamie L. Manson, Simon Parsons,* Euan K. Brechin,* and Stephen Hill* ((Dedication----optional))

The application of pressure in the study of molecule-based materials has gained considerable recent interest, in part due to their high compressibilities, but also because the relevant electronic/magnetic (low energy) degrees of freedom in such materials are often very sensitive to pressure.[1-11] For example, small changes in the coordination environment around a magnetic transition metal ion can produce quite dramatic variations in both the on-site spin-orbit coupling as well as the exchange interactions between such ions when assembled into 3D networks.[2-5] However, perhaps the most compelling reason to use pressure as a tool for understanding magneto-structural correlations is the possibility of focusing investigations on a single molecule or material, as opposed to using chemical means to influence the coordination environment around a metal center, e.g., by studying families of seemingly related complexes that vary only in the identity of the coordinating ligands.

The latter approach obviously suffers from the ‘non-innocence’ of the ligand, particularly in the solid state. The desire to study increasingly complex materials under pressure has spurred the development of sophisticated spectroscopic tools that can be integrated with high-pressure instrumentation.[12-15] The study of magneto-structural correlations requires not only precise crystallographic data, but also detailed spectroscopic information concerning the unpaired electron(s) that give rise to the magnetic properties. Here, we separately employ X-ray and high-frequency EPR spectroscopies to obtain high-resolution structural and magnetic data from oriented single-crystal samples subjected to pressures of up to 3.5 GPa. We focus on the magnetic coordination polymer [CuF2(H2O)2(pyz)] (1, pyz = pyrazine), which previous powder diffraction studies have shown to undergo successive pressure-induced structural transitions, both of which are believed to involve dramatic reorientations of the Jahn-Teller (JT) axes associated with the CuII ions.[9] Importantly, magnetic susceptibility studies reveal a pronounced change in the effective dimensionality of the extended Cu···Cu exchange interactions (from 2D to 1D) at the first of these transitions. [9] In the present investigation, EPR measurements provide direct information on the disposition of the magnetic dx2-y2 orbital upon entering the first of the two high-pressure phases. Furthermore, while the previous powder X-ray studies established the onset of the phase transitions and the reorientations of the JT axes, we describe the results of single-crystal diffraction experiments which not only yield structural data of substantially higher precision, but which also reveal the way in which relief of strain built-up in compressed H-bonds drives the system through successive phase transitions as pressure is increased. We also identify for the first time an additional high pressure phase characterized by a dimerization of 1D chains. Under ambient conditions complex 1 crystallizes in the monoclinic space group P21/c, with one copper center in the asymmetric unit (Table S1). The metal ion resides in a distorted octahedral environment formed by two O-atoms from the two water molecules, two pyrazine N-atoms and two fluorides, as illustrated in Figure 1. The elongated JT axis lies along the N-Cu-N vector with the two symmetry equivalent bonds measuring 2.454(6) Å; the shorter Cu-O and Cu-F distances are 1.984(4) Å and 1.908(4) Å, respectively (Table 1). The CuII ions are linked into 1D chains along the crystallographic a-axis via the pyrazine. They are also linked into a 2D network within the bc-plane by a series of short Cu-OH···F-Cu hydrogen bonds [2.623(4) Å and 2.607(4) Å (Figure 1)]. The magnetic dx2-y2 orbital associated with the axially elongated CuII ion is expected to lie within the CuF2O2 (bc-) plane, thereby adding significance to the hydrogen bonding interactions. Indeed, extensive experimental studies of 1 at ambient pressures reveal a low

[] Dr A. Prescimone, Prof. S. Parsons, Prof. E. K. Brechin, EaStCHEM School of Chemistry, The University of Edinburgh, West Mains Road, Edinburgh, EH9 3JJ, UKE-mail: ebrechin@staffmail.ed.ac.uk; s.parsons@ed.ac.uk

Ms C. Morien, Dr. S. Tozer, Prof. S. Hill, National High Magnetic Field Laboratory and Department of PhysicsFlorida State University, Tallahassee, FL 32310, USA E-mail: shill@magnet.fsu.edu

Dr D. Allan, Diamond Light Source, Harwell, Science Innovation Campus, Chilton, Oxfordshire, OX11 ODE, UK

Dr J. A. Schlueter, Materials Science Division, Argonne National Laboratory, Argonne, IL 60439, USA

Dr J. L. Manson, Department of Chemistry and Biochemistry, Eastern Washington University, Cheney, WA 99004, USA

[] We acknowledge the support of the EPSRC (UK) and NSF (CHE-0924374). We thank Diamond Light Source for access to beamline I19 (proposal number MT6146) that contributed to the results presented herein. Work at Argonne, a U.S. Department of Energy Office of Science laboratory, is operated under Contract No. DE-AC02-06CH11357. Work performed at the NHMFL is supported by the NSF (DMR-0654118) and the State of Florida. JLM gratefully acknowledges support of the NSF through grant No. DMR-1005825. ST acknowledges support from DOE NNSA DE-FG52-10NA29659.

Supporting information for this article is available on the WWW under http://www.angewandte.org or from

1

temperature transition to an antiferromagnetic state (TN = 2.54 K) with strong 2D character, indicating appreciable Cu···Cu exchange within the bc-plane, with considerably weaker interactions along the chains.[16] Under the application of external hydrostatic pressure the first, and perhaps most obvious, effect is the compression of the unit cell (Table S1). At 1.2 GPa the Cu-X (X = N, F, O) coordination bond distances remain essentially unchanged, but a contraction in both hydrogen bonds to 2.515(13) Å is observed; these distances are amongst the shortest M-OH2···F-M H-bonds to have been observed, pinpointing the build-up of strain in the compressed intermolecular contacts.

Figure 1. Visualization of the 2D H-bonded network with the [O-H···F]2 unit in the crystallographic bc plane in phase I (top) and in phase IV (bottom). Color scheme: Cu = orange, O = red, F = yellow, N = blue, C = gray and H = salmon. The dashed red lines indicate the OH···F H-bonds.

Upon further increasing the pressure to 1.8 GPa, the expected phase modification occurs: the monoclinic symmetry is retained, but the Cu-N bonds are rather dramatically compressed (by 0.4 Å) to a value of 2.039(3) Å, while the Cu-O bonds elongate (by 0.3 Å) to a value of 2.316(3) Å. This is a clear indication that the JT axis has reoriented from the N-Cu-N vector to the O-Cu-O vector. This structural reorganization relieves the tension in the OH···F hydrogen bonds, which increase to 2.702(3) Å and 2.626(3) Å. Previous high-pressure studies performed on powder samples have reported this phase change to be reversible.[9,10] However, the single-crystal sample becomes polycrystalline upon release of the pressure. Moreover, the first phase modification is not observed until substantially higher pressures are achieved in the present investigation (~1.8 GPa as opposed to 1.0 GPa[9] – see also the EPR data below). These differences may be due to the different pressure transmitting media used for the various experiments, and possibly

also due to the different nature of the samples (fine powder versus the more rigid/fragile single-crystals). We note also that the pressures were calibrated at very different temperatures using different manometers.

Table 1. Bond distances (Å) in 1 as function of pressure.

P/GPa Cu1-F Cu1-O Cu1-N0 1.908(4) 1.984(4) 2.454(6)0.5 1.904(3) 1.975(4) 2.430(4)0.9 1.907(2) 1.968(3) 2.417(3)1.2 2.030(14) 1.987(6) 2.441(7)1.8 1.898(2) 2.316(3) 2.039(3)2.2 1.904(2) 2.331(3) 2.036(3)2.5 1.906(2) 2.324(3) 2.031(2)2.85 1.907(2) 2.320(3) 2.027(2)3.3 1.886(15) 2.36(2) 2.046(13)

Cu2-F' Cu2-F Cu2-F˝3.3 2.283(12) 1.91(2) 1.91(2)

Cu2-O Cu2-N Cu2-N'3.3 2.22(2) 2.069(13) 2.016(13)

Use of single-crystal diffraction data enables location of the H-atom positions from Fourier difference maps.[17] The mean Cu-O-H angle in phase-II decreases from 108.3(9)° at 1.8 GPa to only 99.7(9)° upon further increasing the pressure to 2.85 GPa (Table S3). A search of the Cambridge Database shows that this almost perpendicular orientation is quite unusual.[18] Nevertheless, it preserves the near linearity in the OH···F contacts as the structure is compressed.

Figure 2. Disposition of the Jahn-Teller axes (higlighted by the dark blue rods) in 1 in the three phases. Color scheme: Cu = orange, O = red, F = yellow, N = blue, C = gray.

Further pressure increments reveal a second structural transition between 2.85 and 3.3 GPa, and this transformation is even more disruptive than the first. The two-fold symmetry is lost and the structure becomes twinned in the triclinic space group P-1 (Table S1). The cause of the symmetry loss lies in the ejection of one water molecule per copper unit from two thirds of the chains, forcing them to dimerize through the F-atoms in order to fill the vacant coordination sites (Figures 1 and 2). Interestingly, the remaining one third of the chains are unchanged, i.e., they have bond lengths and angles equivalent to those seen prior to the 2nd phase transition. Consequently, the structure now contains two different chain types within the asymmetric unit. The structure is also twinned via a pseudo two-fold rotation about the b-axis. The expelled water molecules remain in the crystal lattice, located between the

2

monomeric and dimeric chains, held in place by the OH···F hydrogen bonding network (Figure S1). The mean O···F distances mediated by H-bonds is now 2.67(2) Å, spanning the range from 2.52(2) – 2.74(2) Å, while the average Cu-O-H angle is now 117.5(5)˚.

Figure 3. Experimental EPR spectra recorded at 10 K as a function of the field orientation within the ab-plane of a single-crystal. The three panels correspond to three different pressures and microwave frequencies (indicated in the figure). Spectra were recorded every 10o, with 0o corresponding to the field along a; the field orientations corresponding to the top and bottom traces are indicated in each panel.

The high-pressure phase obtained here for a single-crystal is different from phase-III described by Halder.[9] The latter phase is monoclinic; the Cu-atoms remain 6-coordinate, but the JT axis shifts to the F-Cu-F direction. The powder pattern calculated from the triclinic structure described here, which we shall call phase-IV, is quite clearly different from the experimental powder data reported by Halder (see Fig. S2 in reference 9). It is possible that small deviations in hydrostaticity that arise from use of polycrystalline or single-crystal samples, or from the different hydrostatic media used in the two studies, are responsible for the differences in ours and Halder’s results. The transition to phase-IV reported here causes major changes in the coordination environment of the CuII ions in the dimerized chains (Table 1): in particular, the JT axes are now oriented along the distorted O-Cu-F bonds (Figure 2), which also undergo a significant reorientation relative to the O-Cu-O JT axes associated with the monomeric chains. Most importantly, adjacent CuII ions within the dimerized chains are now bridged directly by F-atoms, likely engendering appreciable exchange coupling between these spins. The hydrogen bonding network is also affected: the Cu-OH···F-Cu pathways between the monomeric and dimerized chains remain relatively unaffected; however, the linkages between dimeric chains become considerably more complex (Fig. S1).The reorientation of the CuII JT axes at the first pressure-induced transition has a profound influence on the magnetic properties of 1. The consequent reorientation of the ground state dx2-y2 orbital into the CuF2N2 plane cuts off the 2D magnetic interactions within the bc-plane, while promoting appreciable 1D correlations along the Cu-pyz-Cu chains. Previous magnetic susceptibility measurements have provided indirect evidence for this JT reorientation on the basis of a change in the dimensionality of the extended magnetic (Cu···Cu) interactions.[9] On the other hand, single-crystal EPR studies can provide direct information on the symmetries of the wave functions associated with the unpaired magnetic electrons through their spin-

orbit mixing with excited ligand field states. This mixing manifests itself in a characteristic anisotropy of the Lande g-tensor. EPR measurements were performed initially at a low pressure of 0.67 GPa. Figure 3(a) displays experimental spectra recorded as a function of field orientation at a frequency of 69.3 GHz and a temperature of 10 K. The crystal was oriented with its c-axis approximately perpendicular to the plane of rotation, i.e., the field was rotated in the ab-plane. A single, sharp peak is observed that displays strong angle dependence with Lande value extrema of g// = 2.42 and g = 2.08 (see Figure 4); the larger value corresponds to the direction parallel to the JT (a-) axis. These values compare extremely well with published results performed under ambient conditions using a commercial X-band (9 GHz) spectrometer.[16]

The next set of measurements were performed at 1.82 GPa, with the 65.7 GHz data displayed in Figure 3(b). This time, two signals are observed: one displays virtually identical angle-dependence to the low-pressure signal (see Figure 4); however, the second signal displays a comparatively weaker angle-dependence, with low g-values in the 2.05 to 2.10 range, suggesting that it corresponds to field rotation within the plane of the dx2-y2 orbital ( orientation). Similar measurements were performed at a frequency of 96 GHz (see Figure 4). These data provide the first hint that some fraction of the sample has transitioned into the second phase in which the JT axis switches to the O-Cu-O bonds. Nevertheless, the transition appears to be incomplete, which might signify a slight non-uniformity of the pressure within the cell. Also evident is a broadening of the EPR signal, particularly in the regions of maximum variation with angle. This indicates strains in the sample, i.e., a degradation of the sample brought about either by thermal cycling or non-hydrostatic pressures within the cell.

Figure 4. EPR peak positions seen in Figure 3 (together with data obtained at other frequencies – not shown), plotted as their corresponding Lande g-factors versus the orientation of the applied field within the ab-plane of the crystal.

Upon further increasing the pressure to 2.1 GPa (Figure 3(c)), the strongly angle-dependent signal (red/black data in Figure 4) vanishes completely, and the remaining EPR intensity is observed in a single, broad low-g signal (blue/green data in Figure 4). This high-pressure EPR signal displays a weak angle-dependence (Figure 4) that agrees remarkably well with bc-plane rotations performed under ambient conditions.[16] However, in the present case, the field was rotated within the ab-plane, thereby providing the most direct evidence that the magnetic dx2-y2 orbital has switched from the

3

CuF2O2 plane to the CuF2N2 plane. Unfortunately, efforts to access the 2nd phase transition were hindered by multiple failures of the plastic pressure cells used for the EPR studies.[15] Future efforts that use diamonds with smaller culets will allow us to reach pressures in excess of 3 GPa more reliably.We return to the 2nd phase transition that occurs at ~3 GPa. As noted above, the dimerized chains involve direct Cu-F-N bridges, thus leading to the formation of Cu2F2 dimers that are coupled along the a-axis via the pyrazine. Magnetically, F-bridges have been shown to behave very much like OH-bridges;[19] the latter have been more widely characterized. In other words, the Cu2F2 dimers should have similar properties to their Cu2(OH)2 counterparts, for which extensive information can be found in the literature. On this basis, the 104o Cu-F-Cu bridging angle would seem to suggest antiferromagnetic coupling; ferromagnetism is typically limited to a narrow angle range between 96o and 98o. Consequently, the dimerized chains can likely be viewed as antiferromagnetic spin ladders with different exchange coupling strength on the rungs and along the rails, the latter determined by the pyrazine bridges. Such chains would very likely possess a spin gap separating a singlet (diamagnetic) ground state from a band of dispersive triplet excited states;[20] the magnitude of this gap would be dependent on the ratio of the two exchange constants (Jrung:Jladder), as well as their magnitudes. Consequently, the dimerized chains would most likely not contribute to the low-temperature magnetic susceptibility or the EPR response. Because the monomer chains are essentially indistinguishable from those in the intermediate pressure phase, it is conceivable that the only magnetic signature of the 2nd phase transformation would be a reduction of the susceptibility (or loss of EPR intensity) by a factor of three due to the dimerization of two thirds of the –Cu–pyz– chains.It should be emphasized that the above discussion of the magnetism within the high-pressure phase is somewhat speculative. Efforts are currently under way aimed at measuring the pressure dependence of the magnetic susceptibility of a crystal in phase-IV in order to bolster our understanding. We also hope that the present investigation may motivate ab-initio calculations aimed at estimating the magnetic exchange interactions within the dimerized spin ladders, thereby providing a means of confirming whether the interaction within the Cu2F2 dimers is indeed antiferromagnetic. Such information could then be used to determine the magnitude of the spin gap associated with the dimerized chains. Finally, we note that high magnetic fields could eventually prove very useful for closing the singlet-triplet gap,[20] thereby ‘switching-on’ the magnetism associated with the dimerized chains.

Received: ((will be filled in by the editorial staff))Published online on ((will be filled in by the editorial staff))

Keywords: High Pressure · EPR · X-ray crystallography · Jahn-Teller reorientation · Cu(II)

[1] D. S. Clarke, S. P. Strong, P. M. Chaikin, E. I. Chashechkina, Science 1998, 279, 2071.

[2] W. Fujita, K. Awaga, Mol. Cryst. Liq. Cryst. Sci. Technol. Sect. A. 2000, 341, 389.

[3] M. Mito, H. Deguchi, T. Tajiri, M. Yamashita, H. Miyasaka, Phys. Rev. B. 2005, 72, 144421.

[4] A. Sieber, R. Bircher, O. Waldmann, G. Carver, G. Chauboussant, H. Muttka, H.-U. Güdel, Angew. Chem. Int. Ed. 2005, 44, 4239.

[5] (a) A. Prescimone, C. J. Milios, J. Sanchez-Benitez, K. V. Kamenev, C. Loose, J. Kortus, S. A. Moggach, M. Murrie, J. E. Warren, A. R. Lennie, S. Parsons, E. K. Brechin, Dalton Trans. 2009, 4858.

[6] K. W. Galloway, S. A. Moggach, P. Parois, A. R. Lennie, J. E. Warren, E. K. Brechin, R. D. Peacock, R. Valiente, J. González, F. Rodriguez, S. Parsons, M. Murrie, CrystEngComm., 2010, 12, 2516.

[7] S. A. Moggach, S. Parsons, Spectrosc. Prop. Inorg. Organomet. Compd. 2009, 40, 324.

[8] P. Parois, S. A. Moggach, J. Sanchez-Benitez, K. V. Kamenev, A. R. Lennie, J. E. Warren, E. K. Brechin, S. Parsons, M. Murrie, Chem. Commun. 2010, 46, 1881.

[9] G. J. Halder, K. W. Chapman, J. A. Schlueter, J. L. Manson, Angew. Chem. Int. Ed. 2010, 49, 419, and references therein.

[10] J. L. Mustfeldt, Z. Liu, S. Li, J. Kang, C. Lee, P. Jena, J. L. Manson, J. A. Schlueter, G. L. Carr, M.-H. Whangbo, Inorg. Chem. 2011, 50, 6347.

[11] A. L. Leitch, K. Lekin, S. M. Winter, L. A. Downie, H. Tsuruda, J. S. Tse, M. Mito, S. Desgreniers, P. A. Dube, S. Zhang, Q. Liu, C. Jin, Y. Ohishi, R. T. Oakley, J. Am. Chem. Soc. 2011, 133, 6051.

[12] L. Merrill, W. A. Bassett, Rev. Sci. Instrum. 1974, 45, 290.[13] A. Dawson, D. R. Allan, S. Parsons, M. Ruf, J. Appl. Cryst., 2004, 37,

410.[14] J. Diederichs, A. K. Gangopadhyay, J. S. Schilling, Phys. Rev. B 1996,

54, R9662.[15] D. E. Graf, R. L. Stillwell, K. M. Purcell, S. W. Tozer, High Pressure

Research 2011, 31, 533.[16] J. L. Manson, M. M. Conner, J. A. Schlueter, A. C. McConnel, H. I.

Southerland, I. Malfant, T. Lancaster, S. J. Blundell, M. L. Brooks, F. L. Pratt, J. Singleton, R. D. McDonald, C. Lee, M.-H. Whangbo, Chem. Mater. 2008, 20, 7408.

[17] P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout, D. J. Watkin, J. Appl. Cryst. 2003, 36, 1487.

[18] F. H. Allen, Acta. Cryst. 2002, B58, 380.[19] S. C. Lee, R. H. Holm, Inorg. Chem. 1993, 32, 4745.[20] T. Giamarchi, A. M. Tsvelik, Phys. Rev. B 1999, 59, 11398.

4

Entry for the Table of Contents (Please choose one layout)

Layout 1:

Controlled Under Pressure X-ray crystallography and high-frequency EPR measurements on [CuF2(H2O)2(pyz)] reveal a series of pronounced structural transitions that involve successive ~90o reorientations of the Jahn-Teller axes associated with the CuII ions. The second transition forces a dimerization involving two thirds of the CuII sites due to ejection of one of the water molecules from the coordination sphere.

Alessandro Prescimone, Chelsey Morien, David Allan, John Schlueter, Stan Tozer, Jamie L. Manson, Simon Parsons,* Euan K. Brechin,* and Stephen Hill*

__________ Page – Page

Pressure-Driven Orbital Reorientations and Coordination Sphere Reconstructions in [CuF2(H2O)2(pyz)]

5

Supporting Information

Pressure-Driven Orbital Reorientations and Coordination Sphere Reconstructions in

[CuF2(H2O)2(pyz)]

Alessandro Prescimone, Chelsey Morien, David R. Allan, John Schlueter, Stan Tozer, Jamie L. Manson,

Simon Parsons, Euan K. Brechin, and Stephen Hill

Table S1. Crystallographic details for the high pressure refinements of 1.

ambient 0.50 GPa 0.90 GPa 1.20 GPa 1.80 GPa

Chemical formula C4H8CuF2N2

O2

C4H8CuF2N2

O2

C4H8CuF2N2

O2

C4H8CuF2N2

O2

C4H8CuF2N2

O2

Mr 217.66 217.66 217.66 217.66 217.66

Crystal system, space group

Monoclinic, P21/c

Monoclinic, P21/c

Monoclinic, P21/c

Monoclinic, P21/c

Monoclinic, P21/c

a, b, c (Å) 7.7004 (2), 7.5648 (3), 6.9033 (2)

7.6300 (6), 7.5027 (14), 6.8168 (12)

7.5857 (14), 7.456 (2), 6.751 (2)

7.6471 (10), 7.551 (5), 6.8191 (10)

6.8372 (3), 7.5981 (9), 7.0280 (8)

, , (°) 90, 111.062 (2), 90

90, 112.657 (8), 90

90, 113.769 (15), 90

90, 112.833 (8), 90

90, 115.095 (6), 90

V (Å3) 375.27 (2) 360.12 (10)

349.44 (17)

362.9 (3) 330.64 (6)

Z 2 2 2 2 2

F(000) 218 218 218 218 218

Radiation type Mo K Mo K Mo K Synchrotron, = 0.48590 Å

Mo K

(mm-1) 2.90 3.02 3.11 3.00 3.29

Crystal size (mm) 0.09 × 0.07 × 0.04

0.07 × 0.05 × 0.04

0.07 × 0.05 × 0.04

0.06 × 0.06 × 0.05

0.07 × 0.05 × 0.04

Diffractometer Bruker Kappa

Bruker Kappa

Bruker Kappa

Area diffractom

Bruker Kappa

6

Apex2 Apex2 Apex2 eter Apex2

Scan method & scans

scans scans scans scans

Tmin, Tmax 0.49, 0.89 0.68, 0.89 0.73, 0.88 0.54, 0.89 0.81, 0.88

measured, independent and observed [I > 2.0(I)] reflections

2543, 990, 715

4080, 338, 315

1984, 335, 294

1119, 231, 187

1888, 318, 273

Rint 0.051 0.022 0.024 0.046 0.024

max (°) 30.2 26.0 26.0 15.7 26.3

R, wR(F2), S 0.056, 0.174, 1.00

0.026, 0.072, 1.03

0.022, 0.062, 0.98

0.046, 0.121, 1.00

0.019, 0.050, 1.01

reflections 964 336 328 221 305

parameters 68 68 68 68 68

No. of restraints 22 22 22 16 22

H-atom treatment refined Refined Refined refined Refined

Weighting scheme Chebychev polynomial

Chebychev polynomial

Chebychev polynomial

Chebychev polynomial

Chebychev polynomial

max, min (e Å-3) 0.95, -1.27 0.27, -0.22 0.20, -0.21 0.32, -0.46 0.19, -0.21

2.20 GPa 2.50 GPa 2.85 GPa 3.30 GPa

Chemical formula C4H8CuF2N2O2 C4H8CuF2N2O2 C4H8CuF2N2O2 C12H24Cu3F6N6O6

Mr 217.66 217.66 217.66 652.99

Crystal system, space group

Monoclinic, P21/c

Monoclinic, P21/c

Monoclinic, P21/c

Triclinic, P¯1

a, b, c (Å) 6.8385(7), 7.5892(12), 7.0257(11)

6.8228(6), 7.5581(10), 7.0043 (9)

6.8098(6), 7.5078(10), 6.9873 (9)

6.7987(8), 10.452 (6), 7.272 (2)

, , (°) 90, 115.204(7), 90

90, 115.566(6), 90

90, 115.987(6), 90

86.24(3), 115.844 (16), 88.71(2)

V (Å3) 329.91 (8) 325.83 (7) 321.12 (7) 463.3 (3)

Z 2 2 2 1

F(000) 218 218 218 327

Radiation type Synchrotron, = 0.48590 Å

Synchrotron, = 0.48590 Å

Synchrotron, = 0.48590 Å

Synchrotron, = 0.48590 Å

(mm-1) 3.30 3.34 3.39 3.52

7

Crystal size (mm) 0.06 × 0.06 × 0.05

0.08 × 0.07 × 0.03

0.08 × 0.07 × 0.03

0.60 × 0.06 × 0.05

Diffractometer Area diffractometer

Area diffractometer

Area diffractometer

Area diffractometer

Scan method scans scans scans scans

Tmin, Tmax 0.68, 0.88 0.81, 0.88 0.80, 0.87 0.70, 0.87

measured, independent and observed [I > 2.0(I)] reflections

1543, 327, 293 1994, 458, 397 1958, 462, 400 2269, 608, 532

Rint 0.034 0.026 0.025 0.041

max (°) 15.6 17.7 17.9 16.1

R, wR(F2), S 0.022, 0.061, 0.98

0.023, 0.063, 0.94

0.023, 0.063, 1.14

0.109, 0.072, 1.15

reflections 316 433 454 532

parameters 68 68 68 77

No. of restraints 22 22 22 13

H-atom treatment geometrical Geometrical Geometrical geometrical

Weighting scheme Chebychev polynomial

Chebychev polynomial

Chebychev polynomial

Chebychev polynomial

max, min (e Å-3) 0.19, -0.20 0.25, -0.21 0.22, -0.22 1.10, -1.11

8

Table S2. H-bond distances and Cu-O-H average angle in 1 as functions of pressure.

Figure S1. Representation of the structure of 1 in phase IV, containing the F-bridged dimerized chains,

the monomeric chains and interstitial water molecules. Color scheme: Cu = orange, O = red, F = yellow,

N = blue, C = grey and H = white.

P/GPa O-H…F O-H…F Cu-O-Haverage

0 2.623(4) 2.607(4) 113.0(9)˚

0.5 2.620(4) 2.599(4) 118.3(9)˚

09 2.610(3) 2.587(3) 108.7(9)˚

1.2 2.515(13) 2.515(13) -

1.8 2.702(3) 2.626(3) 108.3(9)˚

2.2 2.694(2) 2.618(3) 103.2(9)˚

2.5 2.693(2) 2.608(2) 100.2(9)˚

2.85 2.691(2) 2.600(2) 99.7(9)˚

3.3 2.72(2) 2.74(2) 117.5(5)˚

2.69(3) 2.52(2)

9

Experimental section

Table S3. Details of resolution, X-ray source and different crystals used as function of pressure.

P/kbar lamda/Ang theta_max/deg d/Ang crystal#0 0.71 30.223 0.71 15 0.71 26.043 0.81 29 0.71 25.947 0.81 212 0.48 15.663 0.89 318 0.71 26.306 0.80 222 0.48 15.644 0.89 325 0.48 17.654 0.79 4

28.5 0.48 17.854 0.78 433 0.48 16.062 0.87 3

X-ray crystallography

High-pressure single-crystal X-ray diffraction experiments were carried out using a Merrill-Bassett

diamond anvil cell (half-opening angle 40o),[1] equipped with Boehler-Almax diamonds with 600 μm

culets and a tungsten gasket.[2] Petroleum ether was used as the hydrostatic medium and a small ruby chip

was loaded into the cell as the pressure calibrant, with the ruby fluorescence used to measure the pressure.

[3] High-pressure data collections were taken at 0.5, 0.90, 1.20, 1.8, 2.2, 2.5, 2.85 and 3.3 GPa all at room

temperature. Structures at 1.20, 2.2, 2.5, 2.85 and 3.3 GPa were collected using synchrotron radiation of

wavelength = 0.4859 Å on a Rigaku diffractometer at Station I19 at the Diamond Light Source,

Harwell Science and Innovation Campus, whilst data at ambient, 0.5, 0.8 and 1.8 GPa were collected on a

Mo-Ka Bruker Smart APEX II diffractometer.[4] Integrations were carried out using the program SAINT[5]

and absorption corrections with the programs SADABS[6] and SHADE.[7] Refinements of the compressed

forms of 1 were carried out starting from the coordinates obtained from the model of the previous

pressure. The program CRYSTALS[8] was used to refine the structures of 1 against F2 using all the

reflections except for the 3.3 GPa dataset that had to be refined on F using reflections with I >2. Due to

the loss of the two-fold symmetry the 3.3 GPa dataset was twinned with the following twin law: (-1 0 0, -

0.194 1 -0.267, 0 0 -1). Unit cells and refinement parameters are reported in Table S1. Structures and

packing diagrams were drawn using DIAMOND.[9]

10

High-frequency EPR

The high-pressure EPR experiments were performed using a miniature plastic turnbuckle diamond anvil

cell (DAC) and nonmetallic gasket described in ref [10]. The dimensions of the DAC (Ø11.5 12.5 mm

long) permitted its insertion into a custom cylindrical microwave cavity made of OFHC copper without

completely suppressing its resonant characteristics, thus allowing cavity perturbation measurements

involving a small single-crystal within the DAC; the cavity perturbation technique is described elsewhere.

[11] Measurements were performed within a split-pair 7 T superconducting magnet, enabling in situ

orientation of the crystal, thereby enabling discrimination of the anisotropic EPR signal due to the sample

from background signals associated with the DAC.[12] A fiber optic for pressure calibration via ruby

fluorescence[3] was inserted into the cavity along the axis of the DAC without adversely affecting the

resonator Q-factor.  Several pressure media were employed for this portion of the study in an attempt to

find the most hydrostatic fluid. Single crystals were freshly cleaved in such a manner to allow for visual

alignment with respect to the applied field to within 5°. A manuscript describing the high pressure, high-

frequency EPR experimental technique is in preparation.[13]

References[1] L. Merrill, W. A. Bassett, Rev. Sci. Instrum. 1974, 45, 290.

[2] S. A. Moggach, D. R. Allan, S. Parsons, J. E. Warren, J. Appl. Cryst. 2008, 41, 249.

[3] G. J. Piermarini, S. Block, J. D. Barnett, R. A. Forman, J. Appl. Phys. 1975, 46, 2774.

[4] Bruker-Nonius APEX-ll, Version V1, Madison, Wisconsin, USA, 2004.

[5] Bruker-Nonius SAINT version 7, Bruker-AXS: Madison, Wisconsin, USA, 2006.

[6] G. M. Sheldrick, SADABS Version 2004-1, Bruker-AXS: Madison, WIsconsin, USA, 2004.

[7] S. Parsons, S.SHADE, The University of Edinburgh: Edinburgh, United Kingdom, 2004.

[8] D. J. Watkin, K. Prout, J. R. Carruthers, P. W. Betteridge, R. I. Cooper, CRYSTALS Issue 12, Chemical Crystallography

Laboratory, University of Oxford: Oxford, UK, 2003.

[9] K. Brandenburg, H. Putz, DIAMOND, Crystal Impact: Bonn, Germany, 2005.

[10] D. E. Graf, R. L. Stillwell, K. M. Purcell, S. W. Tozer, High Pressure Research 2011, 31, 533.

[11] M. Mola, S. Hill, P. Goy, M. Gross, Rev. Sci. Inst. 2000, 71, 186.

11

[12] S. Takahashi, S. Hill, Rev. Sci. Inst. 2005, 76, 023114.

[13] C. C. Beedle, C. Morien, S. Tozer, S. Hill, in preparation.

12