High-Pressure Crystallography Francesca P. A. Fabbiani Emmy-Noether Jr. Research Group Welcome Part 1 • Introduction to high-pressure research • Experimental setup Dr. Michael Ruf Product Manager, SC-XRD Madison, WI, USA Dr. Francesca P. A. Fabbiani High-pressure crystallography with emphasis on single-crystal X-ray structure determination High-pressure crystallization Solid-state polymorphism Molecular compounds with emphasis on pharmaceuticals and biomolecules Part 2 • Collecting high-pressure data, solving and refining structures • Molecular organic materials, single crystals Introduction to High-Pressure Research High-Pressure Research Range of pressures and temperatures now accessible with static compression techniques in the laboratory. • Planetary science and physics (e.g. minerals, perovskites, clathrates, ices – N.B. 15 polymorphs of H2O to date!) • Synthesis of novel materials (e.g. superhard materials, nanoporous materials, MOFs) Mao H , Hemley R J PNAS 2007;104:9114-9115 ©2007 by National Academy of Sciences • Chemistry/molecular compounds (e.g. amino acids, molecular magnets, pharmaceuticals) • Proteins, pressure-induced denaturation, folding Phase diagram of water Source: Martin Chaplin, http://www.lsbu.ac.uk/water/phase.html High-Pressure Research Cubic BN • Industrial applications, e.g. pressure treatment of food • Origin of life/ sustained life at the bottom of the oceans Source: Wikimedia Commons • Access new materials, probe intermolecular interactions • Understand polymorphic transformations Beta sheet Paracetamol Source: Wikimedia Commons • Interplay with theory Source: Wikimedia Commons High-Pressure Scale (1000 atmospheres ≈ 1 kbar = 100 MPa = 0.1 GPa) Non-equilibrium pressure of H2 Pressure gas in intergalactic inside a light bulb space Hydrogen becomes metallic Centre of the Earth Peak pressure of fist on Ice skater concrete standing on ice during karate strike Centre of the Sun GPa 10-36 10-16 10-5 10-4 10-2 10-1 1 10 102 103 104 Greatest Graphite Sea level Earth depth in becomes atmosphere at oceans diamond altitude of 300 miles Molecular materials 108 Centre of Jupiter Orders of magnitude available for pressure variation >> temperature variation Experimental Setup Pressure Generation for in situ XRD: Diamond-Anvil Cell 1958 First diamond anvil cell at NIST Gaithersburg Museum 2009 Aluminium DAC mounted on a goniometer head 5 cm DAC manufactured by Dr. H. Ahsbahs F. P. A. Fabbiani et al. CrystEngComm (2010), 12, 2354-2360 Source: G. J. Piermarini, Wikimedia Commons Pressure Generation for in situ XRD: Diamond-Anvil Cell F P= F/A (kgms-2/m2 = N/m-2 = Pa) Tightening screws Diamond Ruby Backing plate Sample Gasket A Diamond F Diamond Choose your DAC carefully: backing plates material, diamond cut, purity and culet size, gasket hole size, DAC opening angle DAC Loading • • • Prepare gasket, i.e. sample chamber Load sample together with pressure calibrant, e.g. ruby chip (fluorescence) or quartz single crystal (diffraction) Sample can be in the solid (direct compression in an hydrostatic medium), liquid or gas phase (in situ crystal growth). A solution can also be loaded for in situ crystal growth Equipment for DAC preparation • EDM machine (spark eroder) or laser to drill gasket hole • Stereomicroscope with good magnification, large working distance and polariser/analyser • Small spectrometer with large working distance to monitor ruby fluorescence if using ruby as pressure calibrant DAC Loading Three main sample loading methods available Sample loading methods depend on: • Nature of the sample under investigation • Aim of experimental investigation The focus here is on molecular organic materials for single-crystal X-ray diffraction 300 µm 300 µm “Fathers” of high-pressure research: G. H. J. A. Tammann (1861-1938), P. W. Bridgman (1882-1961) DAC Loading Approach 1: Direct compression in a hydrostatic medium • Good for studying polymorphism in small molecules • Less effective for studying polymorphism in larger and/or rigid molecules: kinetic barrier associated with molecular rearrangement is usually large • Good for studying evolution of structure as a function of pressure, for obtaining p-T phase diagrams and isothermal equation of states • Choice of hydrostatic medium: solubility/ freezing pressure considerations DAC Loading Approach 2: In situ crystallisation and growing from the melt • Excellent method for crystallising new polymorphs of compounds with melting points < 40°C and for comparing with low-temperature structures • Not effective for higher melting organic compounds, which can decompose before the onset of melting p L S V T DAC Loading Approach 2: In situ crystallisation and growing from the melt heat cool DAC Loading Approach 3: In situ crystallisation from solution • Excellent method for crystallising new polymorphs and solvates. No limitation to low melting point or small molecules • Can vary solvent, pressure, temperature and concentration • Prerequisite: relatively high solubility of the solute; solubility of solute increases with increasing temperature and decreases with increasing pressure Outline • Introduction to high-pressure research • Experimental setup Part 1 Part 2 • Collecting high-pressure data, solving and refining structures Molecular organic materials, single crystals Collecting high-pressure data, solving and refining structures Data Collection • • Centre on the diffractometer, 2-step procedure: optical centring and direct beam centring (Dawson et al. 2004) or diffractometric measurements (King & Finger 1979, Dera & Katrusiak 1999) Choose suitable data collection strategy (and wavelength, if applicable) and exposure time; maximise no. frames per run (this helps during data integration if integrating with Bruker software) A. Dawson et al. J. Appl. Cryst. (2004), 37, 410-416 H. E. King & L. W. Finger J. Appl. Cryst. (1979), 12, 374-378 P. Dera & A. Katrusiak J. Appl. Cryst. (1999), 32, 510-515 Set up @ GZG Göttingen ~ 144 mm Short collimator and long beamstop Diffracted beam D ΨD 2θ Data collection in transmission mode Incident beam I Collimator Axis of symmetry Data Collection For one orientation of the DAC, the accessible region of reciprocal space is determined by the detector distance and the DAC opening angle ω Steel support of the DAC starting to obscure the detector R. J. Angel et al. Phase Transitions (1992), 39, 13-32 By a combination of ω-scans using different orientations of the DAC in φ, and different orientations of the detector in 2θ, about ⅓ of all reflections can be collected (Angel et al. 1992). This can be increased by collecting more data with a different orientation of the DAC with respect to χ (see later) Data Collection Example of data collection strategy for a 3-circle diffractometer, detector distance = 7 cm and DAC perpendicular to the beam @ phi = 0°, ½ DAC opening angle = 45° Scan Type 2 Theta Omega Phi Sweep Scan direction Omega -28 -40 0 30 +ve Omega 28 40 0 65 -ve Omega -28 -220 0 65 +ve Omega 28 -140 0 30 -ve Omega -28 -40 180 30 +ve Omega 28 40 180 65 -ve Omega -28 -220 180 65 +ve Omega 28 -140 180 30 -ve Phi 0 0 60 60 +ve Phi 0 180 60 60 +ve Phi 0 0 150 60 +ve Phi 0 180 150 60 +ve Important: check hardware limits; check diffraction limits and adjust 2 Theta and Omega accordingly Data Collection Set up @ GZG Göttingen X-ray Crystallography Diffracted beam D ΨD 2θ Limited sampling of reciprocal space data completeness and resolution Incident beam I Collimator Absorption and shadowing Axis of symmetry Sample scattering power (and size) Background data collection data processing structure solution structure refinement Data Completeness Coverage of reciprocal space Increasing Data Completeness Collect Rotate by 120° and collect Rotate by 120° and collect On a 3-circle diffractometer • Rotation of the DAC • DAC with large opening angle + non-diffracting backing plates • Data collection with short-wavelength radiation (see later) • Careful orientation of the crystal in the DAC (ambient p) • Multiple or twinned crystals F. P. A. Fabbiani et al. CrystEngComm (2010), 12, 2354-2360 Data Processing • Data indexing identify reflections arising from sample Harvesting reflections If large background variations, reduce the number of images and runs Choose an appropriate value It is useful to exclude certain regions, e.g. Be rings, gasket rings; if sample reflections are scarce and indexing is difficult, try omitting high-resolution regions, where diamond reflections are more abundant. This can also be achieved through a reciprocal lattice viewer (recommended) Data Processing • Data indexing identify reflections arising from sample Reciprocal lattice viewer The reciprocal lattice viewer is an invaluable tool for indexing, for assessing data quality and for twin-spotting. Reflections can be conveniently assigned to different groups and exported for indexing with external programs, e.g. CELL_NOW Data Processing • Data indexing identify reflections arising from sample • Data integration mask out regions of detector obscured by the DAC; choose appropriate resolution; background correction The following are recommendations based on personal experience. There is no “one-fits-all” strategy that will work for every sample: try different options to optimise your integration. Once the integration parameters have been optimised, I would strongly recommend performing successive integration cycles (“UB matrix update”) for best intensities and unit cell parameters Data Processing • Data indexing identify reflections arising from sample • Data integration mask out regions of detector obscured by the DAC; choose appropriate resolution; background correction Try to keep the box size small. If problems with convergence: uncheck this option If background is jumpy choose high frequency; otherwise reduce Twins can be easily handled Data Processing • Data indexing identify reflections arising from sample • Data integration mask out regions of detector obscured by the DAC; choose appropriate resolution; background correction This option might be useful for synchrotron data Threshold for strong reflections: lower this to, e.g. 8 for weak data For using dynamic masks generated with an external program Data Processing Typical frames from CCD Area Detector Diamond reflection Sample reflection Powder ring from Beryllium backing discs (now almost obsolete) Powder ring from steel gasket almost invisible at 2θ = 0 (gasket and beam size dependent) Dynamic mask for integration Shading from DAC opening angle Dynamic masks: A. Dawson et al. J. Appl. Cryst. (2004), 37, 410–416; N. Casati et al. J. Appl. Cryst. (2004), 40, 620-630 Data Processing • Data indexing identify reflections arising from sample • Data integration mask out regions of detector obscured by the DAC; choose appropriate resolution; background correction Generate dynamic masks “on the fly”, e.g. with the Bruker SAINT integration software, V8.07A run from the command line (“Advanced options”) Input DAC geometry Data Processing • Data indexing identify reflections arising from sample • Data integration mask out regions of detector obscured by the DAC; choose appropriate resolution; background correction • Scaling and absorption correction 2-stage procedure: analytical correction for DAC components and gasket shadowing, see programs by S. Parsons, A. Katrusiak and R. J. Angel; multiscan correction to correct for other systematic errors and for scaling, e.g. SADABS. Beware of outliers, e.g. diamond reflections! • Space group determination difficulty related to completeness, redundancy, resolution and crystal orientation systematic absences are not always present Data Processing • Structure solution direct methods, global optimisation methods (borrowed from powder diffraction), molecular replacement, etc.: numerous programs available! • Data merging crucial step; robust-resistant and experimental (1/σ2) weighting scheme with SORTAV SORTAV: R. H. Blessing J. Appl. Cryst. (1995), 30, 421–426 Refinement • Very high-quality high-pressure data can be collected nowadays. It is nevertheless important to be realistic during refinement. Refinement of ADPs for all non-H atoms might not be possible • Most commonly encountered problem: low data to parameter ratio; restraints are your friends: treat them well and be generous • Always investigate outliers before omitting reflections: go back to the original frames • The following are examples taken from my own research CRYSTALS: P. W. Betteridge et al. J. Appl. Cryst. (2003), 36, 1487 SHELXL: G. M. Sheldrick Acta Cryst. (2008), A64, 112-122 Refinement Example of problematic data, lab source In situ crystallisation study Cell setting, space group Monoclinic, P21/n a, b, c (Å) 7.630(2) 17.209(3) 7.3708(11) β (°) 103.923(8) Z 4 Multi-scan abs. correction Tmin/Tmax 0.29 No. of measured, independent and observed [F > 4σ(F)] reflections 801, 203, 172 Rint 0.05 No. of parameters 14 R1[F > 4σ(F)], wR2(F2, all reflections) 0.134, 0216 (sinθ/λ)max (Å-1) and completeness (%) 0.5, 24.4 Constraints: 2 rigid bodies 1 isotropic parameter C9H13NO3 Structure solution: DASH Refinement program: CRYSTALS Refinement Example of good data, lab source In situ crystallisation study Water solvent Tungsten gasket Ruby chip Single crystal 300 µm Cell setting, space group Monoclinic, P21/c a, b, c (Å) 8.9537(11) 5.4541(6) 13.610(4) β (°) 104.93(2) Z 4 Multi-scan abs. correction Tmin/Tmax 0.61 No. of measured, independent and observed [F > 4σ(F)] reflections 3718, 470, 359 Rint 0.08 No. of parameters and restraints 92, 83 R1[F > 4σ(F)], wR2(F2, all reflections) 0.053, 0.103 (sinθ/λ)max (Å-1) and completeness (%) 0.63, 34.4 SIMU, DELU, DFIX [for (N)H positions] restraints C6H10N2O2 Structure solution: Sir92 Refinement program: CRYSTALS Refinement Example of good data, lab source Water solvent Tungsten gasket Ruby chip Single crystal 300 µm Here would expect diamond overlaps Fo Overlap with gasket Shaded reflections Fc Increasing data quality – Part I Synchrotron radiation Useful properties of synchrotron radiation: • • • Brilliance gain in diffracted intensity compared to a lab source, i.e. increase in resolution and completeness Tuneable wavelength: shortwavelength radiation is accessible less absorption and significant gain in completeness Small source size: microfocussing is possible very small samples can be investigated; reduction/elimination of gasket diffraction S k′ 2θ k 1/λ θ d 0 nλ = 2d sin θ Refinement Example of good data, synchrotron radiation In situ crystallisation study 300 µm SIMU, DELU restraints All H-atoms could be located on difference Fourier maps Cell setting, space group Triclinic, P-1 a, b, c (Å) 6.7906(12) 7.3159(4) 15.8428(14) α, β, γ (°) 86.297(6) 78.924(11) 72.713(6) Z 2 Multi-scan abs. correction Tmin/Tmax 0.63 No. of measured, independent and observed [F > 4σ(F)] reflections 6429, 1533, 1294 Rint 0.05 No. of parameters and restraints 172, 100 R1[F > 4σ(F)], wR2(F2, all reflections) 0.045, 0.130 (sinθ/λ)max (Å-1) and completeness (%) 0.62, 52 C9H17NO2 . 7(H2O) Structure solution: Sir92 Refinement program: CRYSTALS Refinement Example of good data on a large molecule, synchrotron radiation Compression study 300 µm High pressure (1.0 GPa) Cell setting, space group Orthorhombic, P212121 a, b, c (Å) 15.9455(4) 21.0511(5) 23.8739(8) Z 4 Multi-scan abs. correction Tmin/Tmax 0.75 No. of measured, independent and observed [F > 4σ(F)] reflections 53048, 10032, 7534 Rint 0.05 No. of parameters and restraints 936, 168 R1[F > 4σ(F)], wR2(F2, all reflections) 0.066, 0.216 (sinθ/λ)max (Å-1) and completeness (%) 0.58, 80 SIMU, DELU, DFIX restraints High pressure (1.0 GPa) C63H88CoN14O14P . 22 H2O Structure solution: (SHELXS) Refinement program: SHELXL Refinement Example of good data on a large molecule, synchrotron radiation Ambient-pressure and temperature study Ambient pressure SIMU, DELU, DFIX restraints Ambient pressure Cell setting, space group Orthorhombic, P212121 a, b, c (Å) 15.8260(9) 22.4438(13) 25.4429(16) Z 4 Multi-scan abs. correction Tmin/Tmax 0.88 No. of measured, independent and observed [F > 4σ(F)] reflections 101570, 26484, 19937 Rint 0.04 No. of parameters and restraints 917, 168 R1[F > 4σ(F)], wR2(F2, all reflections) 0.073, 0.225 (sinθ/λ)max (Å-1) and completeness (%) 0.72, 96 C63H88CoN14O14P . 23.5 H2O Structure solution: (SHELXS) Refinement program: SHELXL Refinement Example of good data on a large molecule, synchrotron radiation Water ordering in channels at high pressure High pressure (1.0 GPa) Ambient pressure Electron density maps generated with shelXle, Fo-Fc @ 0.31 e-/Å3, Fo @ 0.98 e-/Å3 ShelXle: C. B. Hübschle et al. J. Appl. Cryst. (2011), 44, 1281-1284 Increasing data quality – Part II Synchrotron radiation Useful properties of synchrotron radiation: • Shorter wavelength less absorption, more diffraction data and smaller diffraction angles In the lab • Ag radiation (0.56087 Å) is now available as an air-cooled 30 W microsource (supplier: Incoatec) increase in data completeness, “cleaner” background S k′ 2θ k 1/λ θ d 0 nλ = 2d sin θ Ag radiation in the lab • Comparative study on gabapentin heptahydrate, Incoatec Ag microsource vs. Mo sealed tube on a Bruker AXS Apex II diffractometer Data statistics Source Ag-IµS Mo-ST Power/ kW 0.03 2.0 Exposure time (s/0.3°) 20 20 <I> 368.8 (64.9) 378.0 (61.0) <I/σ> 19.6 (3.2) 18.3 (4.7) Unique data 866 (170) 721 (135) <Redundancy> 1.5 (0.9) 1.1 (0.7) <Completeness>/% 40.6 (28.9) 33.7 (22.6) Rint 0.0306 0.0342 R1 (I<2σ(I)) 0.0487 630 refl 0.0532 523 refl. wR2 0.1025 860 refl. 0.232 705 refl. Number in parenthesis refer to the highest resolution shell (1.00 – 0.90 Å) Fore more information: http://www.incoatec.de/?id=101 Ag-IµS, 90 µm beam 300 µm Mo-sealed tube, 500 µm beam Further Reading Further Reading Acknowledgments High-Pressure Crystallography • Prof. Simon Parsons (Edinburgh) • Dr. Heidrun Sowa (Göttingen) • Dr. Jürgen Graf (Incoatec) • Dr. Michael Ruf (Bruker) Funding Question & Answer Questions? • Type any questions you may have into the Q&A panel, then click Send. Thank you for joining! • A PDF copy of the slides and a link to the recording of this webinar will be emailed to you. • For more Bruker SC-XRD webinars, visit http://www.bruker-axs.com/ webinars_sc-xrd.html