Upgrade to Pro — share decks privately, control downloads, hide ads and more …

A challenging EXAFS analysis problem

A challenging EXAFS analysis problem

This talk accompanies a demonstration of the analysis strategy for the work explained in this paper:
http://dx.doi.org/10.1016/j.radphyschem.2009.05.024
This difficult problem combined interpretation of XANES and EXAFS and discusses how to deal with severe limits on information content by application of interesting constraints on the fit.

Bruce Ravel

May 16, 2014
Tweet

More Decks by Bruce Ravel

Other Decks in Science

Transcript

  1. Metal sensors Experiment DNA Model building The fit Post mortem

    A challenging EXAFS analysis problem Bruce Ravel Synchrotron Science Group, Materials Measurement Science Division Materials Measurement Laboratory National Institute of Standards and Technology & Local Contact, Beamline X23A2 National Synchrotron Light Source ASEAN Workshop on X-ray Absorption Spectroscopy Synchrotron Light Research Institute June 2–4, 2014 1 / 37 A challenging EXAFS analysis problem
  2. Metal sensors Experiment DNA Model building The fit Post mortem

    Copyright This document is copyright c 2010-2014 Bruce Ravel. This work is licensed under the Creative Commons Attribution-ShareAlike License. To view a copy of this license, visit http://creativecommons.org/licenses/by-sa/3.0/ or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA. You are free: to Share — to copy, distribute, and transmit the work to Remix — to adapt the work to make commercial use of the work Under the following conditions: Attribution – You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work). Share Alike – If you alter, transform, or build upon this work, you may distribute the resulting work only under the same, similar or a compatible license. With the understanidng that: Waiver – Any of the above conditions can be waived if you get permission from the copyright holder. Public Domain – Where the work or any of its elements is in the public domain under applicable law, that status is in no way affected by the license. Other Rights – In no way are any of the following rights affected by the license: Your fair dealing or fair use rights, or other applicable copyright exceptions and limitations; The author’s moral rights; Rights other persons may have either in the work itself or in how the work is used, such as publicity or privacy rights. Notice – For any reuse or distribution, you must make clear to others the license terms of this work. This is a human-readable summary of the Legal Code (the full license). 2 / 37 A challenging EXAFS analysis problem
  3. Metal sensors Experiment DNA Model building The fit Post mortem

    Transport of metal contaminants in the environment There are numerous natural and man-made point sources of toxic metals which find their way into water systems used for human and agricultural applications. The safe use of water requires monitoring and eventual remediation of bioavailable metal species. 3 / 37 A challenging EXAFS analysis problem image from http://lightsources.org, Credit: Argonne National Laboratory
  4. Metal sensors Experiment DNA Model building The fit Post mortem

    Real-time, field-ready sensors Sophisticated laboratory and synchrotron methods exist to detect and speciate water contaminants at very low concentrations. The real-world task of environmental monitoring requires a fast, flexible, sensitive, selective method of detecting contaminants in the field. Fast Obtain results while still in the field Flexible Easy to carry and easy to use in the field Sensitive Detect contaminant concentrations below regulated human health hazard levels Selective Respond to the target metal but not to other metals 4 / 37 A challenging EXAFS analysis problem
  5. Metal sensors Experiment DNA Model building The fit Post mortem

    Real-time, field-ready sensors Sophisticated laboratory and synchrotron methods exist to detect and speciate water contaminants at very low concentrations. The real-world task of environmental monitoring requires a fast, flexible, sensitive, selective method of detecting contaminants in the field. We want Spock’s tricorder! Fast Obtain results while still in the field Flexible Easy to carry and easy to use in the field Sensitive Detect contaminant concentrations below regulated human health hazard levels Selective Respond to the target metal but not to other metals 4 / 37 A challenging EXAFS analysis problem
  6. Metal sensors Experiment DNA Model building The fit Post mortem

    Catalytic DNA-based sensors The sensor has a receptor, a cleavage site, and paired fluorophore and quencher. 5 / 37 A challenging EXAFS analysis problem J. Liu, et al. A catalytic beacon sensor for uranium with parts-per-trillion sensitivity and millionfold selectivity PNAS, 104:7 (2007) 2056-2061 DOI: 10.1073/pnas.0607875104
  7. Metal sensors Experiment DNA Model building The fit Post mortem

    Building a sensor device These DNA sensors can be incorporated into a hand-held device. Water is dropped onto an array of sensors and read using photodiodes. 6 / 37 A challenging EXAFS analysis problem
  8. Metal sensors Experiment DNA Model building The fit Post mortem

    DNA-based Hg sensor U.S. EPA limit on Hg in water is 10 nM (2 ppb) The DNA-based sensor for Hg has a detection limit of 2.4 nM Questions: How and where does the metal bind? Under what conditions does the metal remain bound to the DNA? How many binding sites are there on a sensor? Do different metals behave differently? Can DNAzymes be designed more rationally? And, of course, what can XAS tell us about any of these questions (keeping in mind the very local nature of the XAS measurement)? 7 / 37 A challenging EXAFS analysis problem J. Liu and Y. Lu. Rational Design of “Turn-On” Allosteric DNAzyme Catalytic Beacons for Aqueous Mercury Ions with Ultrahigh Sensitivity and Selectivity, Angew. Chemie, 46:40 (2007) 7587–7590 DOI: 10.1002/anie.200702006
  9. Metal sensors Experiment DNA Model building The fit Post mortem

    XAS measurements Solutions: 50 mM cacodylic acid as a buffer 100 mM NaClO4 to maintain pH=6.10 glycerol to promote glassification upon freezing Samples: Control 15 mM Hg Sample 3 mM Hg with 3 mM DNA Sample with excess Hg 6 mM Hg with 3 mM DNA Measure EXAFS at 10 K 8 / 37 A challenging EXAFS analysis problem
  10. Metal sensors Experiment DNA Model building The fit Post mortem

    Cryostat Displex cryostat at APS 20BM. He exchange gas 10 mm wide opening for beam ∼12 mm wide inner shroud Fluorescence measured through hole on side with a Ge detector At that time, 20BM did not have a focusing mirror 9 / 37 A challenging EXAFS analysis problem
  11. Metal sensors Experiment DNA Model building The fit Post mortem

    Unforced error #1 Here is the fluorescence spectrum: The Hg Lα peak is the tiny thing near the green line. The neighboring peak is vastly larger! 10 / 37 A challenging EXAFS analysis problem
  12. Metal sensors Experiment DNA Model building The fit Post mortem

    Unforced error #1 Here is the fluorescence spectrum: The Hg Lα peak is the tiny thing near the green line. The neighboring peak is vastly larger! What’s cacodylic acid? 10 / 37 A challenging EXAFS analysis problem
  13. Metal sensors Experiment DNA Model building The fit Post mortem

    Unforced error #1 Here is the fluorescence spectrum: The Hg Lα peak is the tiny thing near the green line. The neighboring peak is vastly larger! What’s cacodylic acid? Wikipedia tells me that cacodylic acid is: The big peak is As Kα (∼10.5 keV), our Hg Lα (∼10 keV) peak is on its shoulder. 10 / 37 A challenging EXAFS analysis problem
  14. Metal sensors Experiment DNA Model building The fit Post mortem

    Unforced error #2 The samples were packaged back at the University of Illinois and were about 15 mm by 3 mm. We had to put the samples in the cryostat upright and slit the beam down to ∼1 mm. 11 / 37 A challenging EXAFS analysis problem
  15. Metal sensors Experiment DNA Model building The fit Post mortem

    Unforced error #2 The samples were packaged back at the University of Illinois and were about 15 mm by 3 mm. We had to put the samples in the cryostat upright and slit the beam down to ∼1 mm. Plan ahead! Forgetting about the details leads to much worse data! 11 / 37 A challenging EXAFS analysis problem
  16. Metal sensors Experiment DNA Model building The fit Post mortem

    Our main sample This poor data is due to low concentration, small beam, and large background from the As. We measured 42 scans, taking about 22 hours. 12 / 37 A challenging EXAFS analysis problem
  17. Metal sensors Experiment DNA Model building The fit Post mortem

    Sample and control Chemistry has certainly happened. The control is clearly Hg in some kind of aqueous form. The sample with DNA is clearly different from the control. 13 / 37 A challenging EXAFS analysis problem
  18. Metal sensors Experiment DNA Model building The fit Post mortem

    First question Is all Hg taken up by the DNA? To answer this, we measured a sample with excess Hg. Let’s go do some linear combination fitting. (Note the isosbestic points.) 14 / 37 A challenging EXAFS analysis problem
  19. Metal sensors Experiment DNA Model building The fit Post mortem

    First question Is all Hg taken up by the DNA? To answer this, we measured a sample with excess Hg. Let’s go do some linear combination fitting. (Note the isosbestic points.) Yes, all the Hg is taken up by the DNA. 47(1)% sample + 53(1)% control 14 / 37 A challenging EXAFS analysis problem
  20. Metal sensors Experiment DNA Model building The fit Post mortem

    The nucleotides O P O H O H O C H H C H O C C C H O H H O H H N C N C H N C N H H C N C H O C N C N H H N H C N C H N C C H O C C C H O H H O H H C H H O P O H O O H O C N C H C C H H H C N H O C H O C C C H H H O H H C H H O P O H O O H O C N C H C H C N H H N C H O C C C H O H H O H H C H H O P O H O O H 15 / 37 A challenging EXAFS analysis problem Adenisine Guanosine Thymidine Cytidine Purines Pyridines
  21. Metal sensors Experiment DNA Model building The fit Post mortem

    2D and 3D representations The 2D figures on the previous page were generated from the canonical SMILES strings: Adenisine C1=NC2=C(C(=N1)N)N=CN2C3C(C(C(O3)COP(=O)(O)O)O)O Thymidine CC1=CN(C(=O)NC1=O)C2CC(C(O2)COP(=O)(O)O)O Guanosine C1=NC2=C(N1C3C(C(C(O3)COP(=O)(O)O)O)O)NC(=NC2=O)N Cytidine C1=CN(C(=O)N=C1N)C2C(C(C(O2)COP(=O)(O)O)O)O Neat! But we need 3D structures to run ... 16 / 37 A challenging EXAFS analysis problem
  22. Metal sensors Experiment DNA Model building The fit Post mortem

    Structure from PubChem http://pubchem.ncbi.nlm.nih.gov/ 17 / 37 A challenging EXAFS analysis problem
  23. Metal sensors Experiment DNA Model building The fit Post mortem

    Cartesian coordinates: 3D SDF file 9700 -OEChem-05141416293D 36 37 0 1 0 0 0 0 0999 V2000 -3.5515 -1.5175 0.1599 P 0 0 0 0 0 0 0 0 0 0 0 0 -0.4389 1.3396 1.0202 O 0 0 0 0 0 0 0 0 0 0 0 0 -0.9101 4.1569 -0.0812 O 0 0 0 0 0 0 0 0 0 0 0 0 -2.7552 -0.1874 0.6247 O 0 0 0 0 0 0 0 0 0 0 0 0 3.6173 1.7470 0.3907 O 0 0 0 0 0 0 0 0 0 0 0 0 3.8378 -2.8022 -0.2452 O 0 0 0 0 0 0 0 0 0 0 0 0 -2.5475 -2.2163 -0.8977 O 0 0 0 0 0 0 0 0 0 0 0 0 -4.7267 -0.9241 -0.7790 O 0 0 0 0 0 0 0 0 0 0 0 0 -4.0197 -2.4002 1.2798 O 0 0 0 0 0 0 0 0 0 0 0 0 1.6113 0.5684 0.1973 N 0 0 0 0 0 0 0 0 0 0 0 0 3.7127 -0.5224 0.0726 N 0 0 0 0 0 0 0 0 0 0 0 0 -1.0101 2.8736 -0.6948 C 0 0 2 0 0 0 0 0 0 0 0 0 -1.5699 1.8660 0.2995 C 0 0 1 0 0 0 0 0 0 0 0 0 0.3733 2.3378 -0.9829 C 0 0 0 0 0 0 0 0 0 0 0 0 0.7701 1.7196 0.3478 C 0 0 1 0 0 0 0 0 0 0 0 0 -2.2796 0.6993 -0.3750 C 0 0 0 0 0 0 0 0 0 0 0 0 1.0112 -0.6708 0.0146 C 0 0 0 0 0 0 0 0 0 0 0 0 3.0176 0.6816 0.2323 C 0 0 0 0 0 0 0 0 0 0 0 0 1.6792 -1.8209 -0.1381 C 0 0 0 0 0 0 0 0 0 0 0 0 3.1656 -1.7831 -0.1119 C 0 0 0 0 0 0 0 0 0 0 0 0 1.0130 -3.1449 -0.3336 C 0 0 0 0 0 0 0 0 0 0 0 0 -1.6278 2.9841 -1.5911 H 0 0 0 0 0 0 0 0 0 0 0 0 -2.2303 2.3332 1.0386 H 0 0 0 0 0 0 0 0 0 0 0 0 (+ several more hydrogen atoms + bonding information) Here is the “SDF” file for thymidine monophosphate from PubChem. Along with lots of stuff not relevant to the EXAFS analysis, we find the Cartesian coordinates of all the atoms in thymidine monophosphate! 18 / 37 A challenging EXAFS analysis problem SDF = Structure data file
  24. Metal sensors Experiment DNA Model building The fit Post mortem

    Cartesian coordinates: Feff input file TITLE Hg decorating thymidine monophosphate HOLE 4 1.0 * Hg L3 edge (12284 eV), S0^2 * mphase,mpath,mfeff,mchi CONTROL 1 1 1 1 PRINT 1 0 0 0 RMAX 6.0 POTENTIALS * ipot Z element 0 50 Hg 1 8 O 2 7 N 3 6 C 4 15 P 5 1 H ATOMS * x y z ipot -3.5515 -1.5175 0.1599 4 -0.4389 1.3396 1.0202 1 -0.9101 4.1569 -0.0812 1 -2.7552 -0.1874 0.6247 1 3.6173 1.7470 0.3907 1 3.8378 -2.8022 -0.2452 1 -2.5475 -2.2163 -0.8977 1 -4.7267 -0.9241 -0.7790 1 -4.0197 -2.4002 1.2798 1 1.6113 0.5684 0.1973 2 3.7127 -0.5224 0.0726 2 -1.0101 2.8736 -0.6948 3 -1.5699 1.8660 0.2995 3 0.3733 2.3378 -0.9829 3 * (and so on...) 1 Do some cutting and pasting 2 Add some boilerplate for the header 3 Make a sensible POTENTIALS list What about the Hg atom? 19 / 37 A challenging EXAFS analysis problem
  25. Metal sensors Experiment DNA Model building The fit Post mortem

    What is the likely location of the Hg atom? 1 Thymine forms its hydrogen bond with adenisine via the N atom 2 The engineered DNA sensor is known to have a T-T mismatch 3 Earlier NMR work was interpreted at having the Hg bridging the T-T mismatch. That said, I don’t know much about this chemistry. 20 / 37 A challenging EXAFS analysis problem Y. Miyake, et al., MercuryII-Mediated Formation of Thymine-HgII-Thymine Base Pairs in DNA Duplexes. J. Am. Chem. Soc. (2006) v.128, 2172-2173 DOI: 10.1021/ja056354d
  26. Metal sensors Experiment DNA Model building The fit Post mortem

    Decorating thymidine with Hg C C C Hg? N C N Hg? C O C O Hg? C C Hg? C OH2 C O Hg? C C O P O− O− Hg? O a b ϕ 21 / 37 A challenging EXAFS analysis problem
  27. Metal sensors Experiment DNA Model building The fit Post mortem

    Hg atom placement, 1 Do a quick first shell fit to determine the Hg 1st shell distance. Not a great fit, but it tells us that the Hg atom is about 2.05 ˚ A away from it’s neighbor. 22 / 37 A challenging EXAFS analysis problem
  28. Metal sensors Experiment DNA Model building The fit Post mortem

    Hg atom placement, 2 Using the known nucleotide structures, I wrote a small program to solve some trigonometry: The Hg atom is ... 1 ... 2.05 ˚ A away from its neighbor 2 ... in the same plane as the neighboring atoms 3 ... equidistant from the second neighbors (6- and 5-member ring options) 4 ... collinear with the 1st and 2nd neighbors (monodentate option) Finally, write out ‘feff.inp’ files with Hg as the absorber. 23 / 37 A challenging EXAFS analysis problem
  29. Metal sensors Experiment DNA Model building The fit Post mortem

    5-member ring option: coordinates TITLE Hg decorating thymidine monophosphate HOLE 4 1.0 * Hg L3 edge (12284 eV), S0^2 CONTROL 1 1 1 1 PRINT 1 0 0 0 RMAX 6.0 POTENTIALS * ipot Z element 0 50 Hg 1 8 O 2 7 N 3 6 C 4 15 P ATOMS * x y z ipot 0.49977 0.63093 2.85314 0 Hg 0.00000 -3.71800 -2.00000 -1.24900 1 O 6.44507 -3.91000 -1.59800 0.29000 4 P 5.56632 -5.35700 -1.58600 0.60000 1 O 6.65531 -3.02000 -2.44600 1.11000 1 O 4.98947 -3.35200 -0.12200 0.45900 1 O 4.59727 -1.95500 0.01100 0.30500 3 C 3.59210 -1.48700 1.40500 0.63700 3 C 3.07534 -0.11000 1.31500 1.03100 1 O 2.05000 -1.52300 2.38600 -0.51700 3 C 4.30462 -1.77700 3.69900 -0.00600 1 O 4.77194 -0.15400 2.19400 -1.16100 3 C 4.35705 0.73400 1.75700 0.00100 3 C 3.07532 1.68600 0.62300 -0.15600 2 N 3.23452 1.54600 -0.35900 -1.10700 3 C 4.21393 0.64900 -0.39000 -1.93000 1 O 4.89315 2.52400 -1.32100 -1.06300 2 N 4.82117 3.58700 -1.40200 -0.18100 3 C 4.78223 4.40700 -2.31400 -0.22400 1 O 5.77995 3.65500 -0.33500 0.78500 3 C 3.89431 4.86400 -0.10800 1.62800 3 C 4.59276 2.71300 0.59300 0.75000 3 C 3.05336 24 / 37 A challenging EXAFS analysis problem
  30. Metal sensors Experiment DNA Model building The fit Post mortem

    5-member ring option: paths Run , drag-n-drop first 6 paths, transfer them to the plotting list, plot in R: This looks sort of promising ... or does it? 25 / 37 A challenging EXAFS analysis problem
  31. Metal sensors Experiment DNA Model building The fit Post mortem

    5-member ring option: VPath We fit a sum of paths to the data, so let’s examine the sum of these paths. In , this is called a “VPath.” Not so promising, after all. 26 / 37 A challenging EXAFS analysis problem
  32. Metal sensors Experiment DNA Model building The fit Post mortem

    Monodentate option: coordinates TITLE Hg decorating thymidine monophosphate HOLE 4 1.0 * Hg L3 edge (12284 eV), S0^2 CONTROL 1 1 1 1 PRINT 1 0 0 0 RMAX 6.0 POTENTIALS * ipot Z element 0 50 Hg 1 8 O 2 7 N 3 6 C 4 15 P ATOMS * x y z ipot 5.74339 -3.80032 -0.29408 0 Hg 0.00000 -3.71800 -2.00000 -1.24900 1 O 9.67837 -3.91000 -1.59800 0.29000 4 P 9.91863 -5.35700 -1.58600 0.60000 1 O 11.35435 -3.02000 -2.44600 1.11000 1 O 8.97789 -3.35200 -0.12200 0.45900 1 O 9.83988 -1.95500 0.01100 0.30500 3 C 8.61105 -1.48700 1.40500 0.63700 3 C 8.95772 -0.11000 1.31500 1.03100 1 O 7.88572 -1.52300 2.38600 -0.51700 3 C 9.54572 -1.77700 3.69900 -0.00600 1 O 10.62446 -0.15400 2.19400 -1.16100 3 C 8.45356 0.73400 1.75700 0.00100 3 C 7.48765 1.68600 0.62300 -0.15600 2 N 6.00394 1.54600 -0.35900 -1.10700 3 C 5.48832 0.64900 -0.39000 -1.93000 1 O 6.34502 2.52400 -1.32100 -1.06300 2 N 4.13555 3.58700 -1.40200 -0.18100 3 C 3.22719 4.40700 -2.31400 -0.22400 1 O 2.05000 3.65500 -0.33500 0.78500 3 C 4.18739 4.86400 -0.10800 1.62800 3 C 4.25452 2.71300 0.59300 0.75000 3 C 5.43826 27 / 37 A challenging EXAFS analysis problem
  33. Metal sensors Experiment DNA Model building The fit Post mortem

    Monodentate option: VPath Same exercise – run feff, drag-n-drop the first few paths, make a VPath, plot with the data. Better than the 5-member ring option, but still not so great. 28 / 37 A challenging EXAFS analysis problem
  34. Metal sensors Experiment DNA Model building The fit Post mortem

    6-member ring option: coordinates TITLE Hg decorating thymidine monophosphate HOLE 4 1.0 * Hg L3 edge (12284 eV), S0^2 CONTROL 1 1 1 1 PRINT 1 0 0 0 RMAX 6.0 POTENTIALS * ipot Z element 0 50 Hg 1 8 O 2 7 N 3 6 C 4 15 P ATOMS * x y z ipot 2.40463 -2.80748 -2.45560 0 Hg 0.00000 -3.71800 -2.00000 -1.24900 1 O 6.29242 -3.91000 -1.59800 0.29000 4 P 6.99112 -5.35700 -1.58600 0.60000 1 O 8.43040 -3.02000 -2.44600 1.11000 1 O 6.50160 -3.35200 -0.12200 0.45900 1 O 6.98896 -1.95500 0.01100 0.30500 3 C 5.87972 -1.48700 1.40500 0.63700 3 C 6.51567 -0.11000 1.31500 1.03100 1 O 5.95606 -1.52300 2.38600 -0.51700 3 C 6.79387 -1.77700 3.69900 -0.00600 1 O 8.11301 -0.15400 2.19400 -1.16100 3 C 5.76519 0.73400 1.75700 0.00100 3 C 5.44613 1.68600 0.62300 -0.15600 2 N 4.19199 1.54600 -0.35900 -1.10700 3 C 2.92421 0.64900 -0.39000 -1.93000 1 O 3.03360 2.52400 -1.32100 -1.06300 2 N 2.05000 3.58700 -1.40200 -0.18100 3 C 2.92356 4.40700 -2.31400 -0.22400 1 O 3.03859 3.65500 -0.33500 0.78500 3 C 4.26357 4.86400 -0.10800 1.62800 3 C 5.47827 2.71300 0.59300 0.75000 3 C 4.68340 29 / 37 A challenging EXAFS analysis problem
  35. Metal sensors Experiment DNA Model building The fit Post mortem

    6-member ring option: VPath Again – run , drag-n-drop the first few paths, make a VPath, plot with the data. I actually like this one quite a bit! The amplitude is off by about a factor of 2, but the phase is quite close. 30 / 37 A challenging EXAFS analysis problem
  36. Metal sensors Experiment DNA Model building The fit Post mortem

    Parameterization Number of independent points k-range: 2 ˚ A−1 to 8.8 ˚ A−1 R-range: 1 ˚ A to 2.8 ˚ A N idp = 2∆k∆R/π ≈ 7.8 1 E0 and amp are variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1,2) 2 Hg-N distance and σ2 are variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3,4) 3 Hg-O distance and σ2 are variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5,6) 4 Assume that the ring is completely rigid, this allows us to approximate the contributions of various single and multiple scattering paths without introducing any more variables. 31 / 37 A challenging EXAFS analysis problem
  37. Metal sensors Experiment DNA Model building The fit Post mortem

    Trigonometry N C N Hg C O C O a b ϕ ϕ =116.25◦ b =1.378 ˚ A a and σ2 Hg·N are variables of the fit. Here’s a formula for a triangle in a plane: D(Hg − C) = a − b cos(θ) cos(ϕ/2) tan(θ) = a + b a − b tan(ϕ/2) Assuming the ring is rigid, then we approximate σ2 Hg·C (and others) by scaling geometrically from σ2 Hg·N 32 / 37 A challenging EXAFS analysis problem
  38. Metal sensors Experiment DNA Model building The fit Post mortem

    Paths Path 1 (SS) N C N Hg C O C O ×1 ∆R1 and σ2 1 are variables Path 2 (SS) N C N Hg C O C O ×2 ∆R2 computed with trigonomtry σ2 2 ∝ σ2 1 Path 3 (SS) N C N Hg C O C O ×2 ∆R3 and σ2 3 are variables Path 4 (MS) N C N Hg C O C O ×4 ∆R4 computed from paths 1 and 2 σ2 4 := σ2 2 Path 5 (MS) N C N Hg C O C O ×2 ∆R5 computed from path 1 σ2 5 := σ2 2 Path 6 (MS) N C N Hg C O C O ×4 ∆R6 computed from paths 1 and 3 σ2 6 := σ2 1 + σ2 3 33 / 37 A challenging EXAFS analysis problem
  39. Metal sensors Experiment DNA Model building The fit Post mortem

    Fitting result amp 1.86 ± 0.44 E0 1.41 ± 1.91 ∆R(N) 0.006 ± 0.028 ∆R(O) −0.058 ± 0.063 σ2(N) 0.0046 ± 0.0045 σ2(O) 0.0096 ± 0.0081 Why is amp near 2? The Hg atom bridges 2 thymines. Our model had Hg bound to 1 thymine. So S2 0 is really 0.93(44)! 34 / 37 A challenging EXAFS analysis problem
  40. Metal sensors Experiment DNA Model building The fit Post mortem

    Uncertainties The data are short – i.e. little information content – and noisy The uncertainties are all quite large, although the best fit values all make sense S2 0 came out right, although with large uncertainty The σ2 approximations are sensible, but certainly not correct The assumption that the ring is rigid is sensible, but certainly not correct The assumption that the Hg atom sits in the plane of the ring is sensible, but certainly not correct Our data are consistent with the Hg atom bound to the N atom in the 6-member nitrogenous base 35 / 37 A challenging EXAFS analysis problem
  41. Metal sensors Experiment DNA Model building The fit Post mortem

    What could we have done better? The As in the cacodylic acid hurt. Use a different buffer. The sample geometry hurt. Use better packaging or a focusing mirror. Those two things could have increased efficiency by about an order of magnitude. Another couple inverse ˚ Angstroms would have made a huge difference! 36 / 37 A challenging EXAFS analysis problem
  42. Metal sensors Experiment DNA Model building The fit Post mortem

    What have we learned? The science question required interpretation of both XANES and EXAFS Quick first shell fit to approximate the first shell distance Made input for from published structural data and a sensible guess for the location of the Hg atom Tried several possible coordination geometries, but only pursued the one that looked promising Dealt with limited information by applying interesting constraints We didn’t exactly solve the structure, but we demonstrated that the EXAFS data are consistent with the assumption from NMR 37 / 37 A challenging EXAFS analysis problem