Work performed with

1
Synchrotrons, Mannitol, Prazosin,
etc. Synchrotron radiation is accessible,
decisive Hydrogen bonding affects molecular
conformation
Work performed with V.V. Chernyshev, A.V.
Yatsenko, O.B. Ryabova, V.A. Makarov, C.E. Botez,
R. Suryanarayanan, C. Nunes. See also poster 14
on R-albuterol, S. Cuffini
2
Physical adsorption of Kr on graphite
X-ray diffraction of Kr on exfoliated graphite.
Crystallography with one Bragg peak - Kr(1,0)
3
Experiments at SSRL in 1979. (Moncton,
Birgeneau, Horn, Brown, PWS) 20x better angular
resolution Substrate coherence length is
2000Å Completely new picture! There is a
disordered phase between the commensurate and
incommensurate solids. Interesting new physics.
Heroic age of synchrotron radiation. Parasitic on
high energy physics. Huge investment of effort
to get one or two weeks of access per year.
4
National Synchrotron Light Source at Brookhaven
National Laboratory
Produces electromagnetic radiation from IR to
?-rays. Easy to obtain access. 75 experimental
stations 2500 users per year Typical of many
facilities worldwide APS, ESRF, SLS, SRS
Daresbury, SSRL, , which are eager for users.
5
Synchrotron access is for Academic Who
publish in the open literature Industrial
Hold data for proprietary reasons The people
who operate these facilities need to have them
widely used!
Access by writing competitive proposals or
arranging collaboration. You have to pay for
the prorated cost of operating the facility -
250/hour 2000/pattern at NSLS.
6
Compare lab vs. synchrotron data sets. This drug
has two polymorphs that cant be quantified
except by Rietveld.
7
THEORY OF POWDER DIFFRACTION
(series of elementary recipes)
CONTENTS OF UNIT CELL
UNIT CELL
RECIPROCAL LATTICE, SPACE GROUP
. . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . .
INTENSITIES
POWDER PEAK POSITIONS
PHYSICAL SAMPLE AFFECTS LINESHAPES
INSTRUMENT AFFECTS PEAK SHAPES
OTHER ARTIFACTS PREFERRED ORIENTATION, ETC.
POWDER DIFFRACTION PATTERN
8
USE OF POWDER DIFFRACITON TO SOLVE A CRYSTAL
STRUCTURE
Chemical knowledge of contents
Data

• Start with the best data you can get (but no
  better).
• Get a list of accurate diffraction peak
  positions.
• Figure out a lattice that explains the peaks.
• Guess the space group (systematic absences,
  molecules).
• Search for the best place to put the molecule(s),
  best conformation of the molecule.
• Refine, refine, refine, refine, refine,
• At any stage, you can be forced to jump back to
  any stage.

9
Dont think that people only use powders for
organic molecules, or that direct space modelling
is the only useful technique Lausenite
Souzalite Fe2(SO4)35H2O (Fe,Mg)3(al,Fe)4(PO4
)4(OH)62H2O J. Majzlan, , PWS A. Le Bail, ,
PWS Direct Methods (EXPO) Real Space (ESPOIR)
10
1, 2 This is a data-driven enterprise.
Students may think that we spend all our time
talking about algorithms, software, etc., but the
results are no better than the data!
Powder diffraction station at X3B1 beamline,
National Synchrotron Light Source, Brookhaven
National Laboratory, U. S. A.
GE (111) analyzer crystal
From storage ring
Scintillation detector
Monochromatic X-ray beam
sample
Ion chamber
Si(111) double monochromator

• Analyzer crystal geometry measures angles
  eliminates significant aberrations of familiar
  Bragg-Brentano diffractometer.
• Capillary sample geometry is very helpful.
  Eliminates preferred orientation, peak shifts
  that bother flat plate

11
3. Indexing Given some values of d spacings,
find a lattice that fits them, i.e., find
A,B,C,D,E,F such that every d can be expressed
as 1/d 2 Ah 2 Bk 2 Cl 2 Dkl Ehl Fhk
for some integers h, k, l. Familiar programs,
in the public domain TREOR, ITO, DICVOL, have
their quirks, but basically they always work,
given sufficiently good data. (Often possible
with good lab diffractometers, nearly always with
synchrotron data.) TOPAS (Alan Coehlo, Bruker
AXS) has indexing tools that are qualitatively
more powerful.
12
Prazosin
Designer drug selective antagonist for
a1-adrenoceptors (blood pressure). Four other
polymorphs claimed in US Patents 4092315,
4739055, 4816455, and JP Patent 03206088.
Department of Medicinal Chemistry, State
Scientific Center of Antibiotics, Moscow, could
not reproduce any of them. Patent literature
Literature Military intelligence Intelligence
13
(No Transcript)
14
Four of the lines in the Powder Diffraction File
for that compound are impurities, not seen in our
pattern. Throw them away and use TOPAS to index
the lab data.
Triclinic TOPAS FOM 15.99 a 8.717Å b
7.572Å c 16.381Å a 90.01 ß 72.43 ?
108.95 Vol 969Å3
x
x
x
x
Errors 0.021 0.024
(There is a lot of not-quite-good data in the
data bases. Is not-quite-good distinguishable
from bad?)
15
Given sufficiently good data -gt pattern can be
indexed easily Data quality sample ?
instrument If a pattern from a good instrument
cannot be indexed, there is something wrong with
the sample
Test your instrument by trying to index known
phases of comparable complexity. Acetaminophen
and Ibuprofen are good organic test cases to get
started.
16
5. Make a model of the molecule, put it into
the lattice. Move the model around seeking best
agreement between calculated and observed
diffraction patterns.
Lots of options software DASH, PSSP, FOX, TOPAS,
PowderSolve, In this case, assumed P1,
coauthor searched nine parameters with software
developed with H. Schenk.
_
Cl
17
6. Refine, refine, If your presumed rough
solution is close enough, you can roll down hill
to the correct solution, using refinement
programs such as GSAS, TOPAS, FULLPROF,
?2
100 refined variables
18
THE GLOBAL TOPOLOGY MAY LOOK MORE LIKE THIS
19
Any fit looks good on this scale
d 1.47Å
?2 2.31, Rwp5.92. No restraints except for
tethering all H atoms.
20
(No Transcript)
21
Same steps for prazosin free base only 6 search
coordinates
Monoclinic, Cc, ?2 2.78, Rwp5.92
22
Planarity of the aromatic rings gives a measure
of the degree of accuracy of the finished atomic
geometry.
23
Prazosin refinements. Geometry of piperazine ring
Bonds Hydrochloride Free Base
N17-C18 1.489Å 1.521Å
C18-C19 1.509Å 1.491Å
C19-N20 1.513Å 1.507Å
N20-C21 1.471Å 1.489Å
C21-C22 1.538Å 1.610Å
C22-N17 1.472Å 1.543Å
Angles
N17 123.1? 117.9?
C18 102.3? 107.7?
C19 114.2? 114.6?
N20 115.8? 111.0?
C21 106.2? 111.0?
C22 105.1? 102.8?
Hydrochloride
N17
C18
Free base
24
Free base 1 N-HN 3.02Å 1 N-HO 3.02Å
Hydrochloride 3 N-HCl
25

• Prazosin conclusions
• That wasnt so hard
• Of relevance to quantitative modeling of
  structure-activity relationships

Hydrochloride Free base Hydrochloride
methanol solvate (single xtl) Prazosin2
tetrachloro-copper(II) (s x)
26

• The rest of the talk
• Enalapril Maleate. Y.H. Kiang, Merck
• Proxy for a real business problem.
• Delta D-Mannitol
• Mannitol hemihydrate

27
Enalapril Maleate is a potent angiotensin
converting enzyme (ACE) inhibitor with two known
polymorphs, Form I and Form II. The single
crystal structure of Form I has been known for
almost twenty years. On the other hand, the
crystal structure of Form II has never been
reported before because of the difficulty to
obtain single crystals of this polymorph, which
is made by water slurry of Form I. The crystal
structure of Form II is of interest for several
reasons 1. Form II is the more stable of the
two polymorphs. 2. The two forms are structurally
similar based on X-ray, IR, and solid-state
NMR. 3. The conformation of ACE inhibitors is
important to their biological activity.
28
Lab(Sealed Tube) and Synchrotron XRD patterns of
Enalapril Maleate
Form I
Cu Ka1
Form II
2 q
Form I
l1.15Å,
Form II
2q
29
Enalapril Maleate
form II
Orthorhombic P212121
a17.838 b6.640 c11.649 b106.29
a33.987 b6.642 c11.210
form I
Monoclinic P21
23 parameters 11 enalapril torsions (2
maleate) 6 orientation 6 position
At the time of the original work, we couldnt
solve from simulated annealing. We could have
benefited from the systematic geometric insights
presented by Claire Gervais
30

Y.-H. Kiang of Merck found the solution by hand,
using Cerius. Y.-H. Kiang, Ashfia Huq, Peter W.
Stephens, Wei Xu, Journal of Pharmaceutical
Sciences 92, 1844-53 (2003)
31
Real business problem _____ has a patented
polymorph of _____ , and suspects that _____ is
selling material that infringes. It is desired
to examine the commercial tablets and determine
the polymorph of the API for potential
litigation. Proxy Examine commercial tablet of
Endocet 500/7.5 Gross tablet 607 mg Acetaminophen
500 mg known lattice structure Oxycodone (as
HCl) 7.5 mg pattern in PDF but lattice
unknown,
In general, Id like to get better info into
the PDF database. Please get in touch if you can
help.
32
of acetaminophen
33
Powder patterns of oxycodone hydrochloride from
ICDD Powder Diffraction File. Strucutures and
lattices are not known.
34
(No Transcript)
35
H H OH OH H H
H - C - C - C - C - C - C - H
OH OH H H OH OH
Back to structures Mannitol
D-Mannitol (naturally produced in various
plants) Long-standing confusion about how many
forms of D-mannitol exist finally laid to rest
by Burger, Henck, co. (2000). (Their II, I, and
III are more commonly known as a, ß, d) a and ß
were solved from single crystals in 1968. d
(lowest melting) identified in 1968, but no
single crystals grown (until 2002, Henck
Benet-Buchholz, unpublished). We solved the
structure of d from a powder sample (with 20
ß) Mannitol is widely used as an excipient in
freeze-drying metastable hydrate discovered by
Lian Yu. TGA shows it is hemihydrate. Structure
solved from powder sample with 26 d, 2 ß, 10
ice.
36
P 21 21 21 8.942 x 18.798 x 4.893Å 205.6 Å3 /
molecule Middle melting
P 21 21 21 8.672 x 16.875 x 5.560Å 203.4 Å3 /
molecule Highest melting
P 21 5.089 x 18.250 x 4.917Å ß 118.304 201.0
Å3 / molecule Lowest melting
Present work
37
All the same steps. At extraction, we did a Le
Bail refinement of the d lattice along with
Rietveld refinement of (known) ß structure.
? 0.70224Å
38
Alpha D-mannitol, H bonds (beta is very similar)
4-cycle
Zig-zag chain
39
Delta D-mannitol, H bonds
One zig-zag chain
40
alpha
beta
The molecules in alpha, beta, delta D-mannitol
are essentially identical
delta
41
Mannitol hemihydrate
Start with the best data possible?
Play the hand youre dealt!
42
Lab x-ray (Minnesota)
Lab x-rays identify the sample TGA -gt hemihydrate
43
Index the peaks that are not any other identified
phase to a triclinic lattice 9.896 x 10.542 x
4.786 Å, 102.59, 86.09, 116.08, 2 mannitol in
P1.
44
Search two independent mannitol molecules, one O
atom. Several different starts.
B
A
45
Hydrogen bonding pattern in hemihydrate
Conformation with one leg lifted
water
Normal conformation
46
Only mannitol hemihydrate has an OH twisted up
into the plane of the C-C-C-C-C-C backbone. What
is the energy cost relative to the conformation
of all other observed mannitol crystal
structures? Is there no way to pack table
mannitols and a water of solvation without
straining the molecule?
47

• Structure determination from powders requires
• Good data
• Fundamental understanding of the available tools
• Motivation
• Choose good problems.
• If you are stuck with a crystallographic problem,
  try a synchrotron. You pay taxes claim your
  share. Find one where there is already a strong
  program of structure determination.
• Also single crystal.