The CIF file, refinement

details and validation of the

structure

CCCW 2011

Artefacts versus errors

Artefacts in crystallography: An error which cannot be

avoided, since it is inherent in the method.

Examples of artefacts:
• Shortened bond distances to light
atoms due to libration

1.04 Å (neutron)

0.96 Å (X ray)

1
 Artefacts versus errors

Artefacts in crystallography: An error which cannot be

avoided, since it is inherent in the method.

Examples of artefacts:
• Shortened bond distances due to libration.
• Shortened bond distances for triple bonds C≡C ou C≡N (electron
density in the bond)
• Wrong hydrogen positions
• Residual electron density around special positions or close to
heavy atoms because of Fourier truncation errors, also called
Fourier truncation ripples.

Fourier truncation errors

The electron density, obtained from the Fourier
transformation, is an (infinite) sum of sinus functions.

ρ ( x , y , z ) = 1V ∑ Fhkl ⋅ e iα ⋅ e − 2πi ( hx + ky + lz )
hkl

hkl

If some of the elements of this sum is missing (finite sum),

because we cannot measure reflections up to very high
angles or because some reflections are not accessible, we
observe “ripples” in the obtained electron density.

http://www.theses.ulaval.ca/2005/23016/apd.html

2
 The errors
An artefact is an error which is unavoidable.

Thus, what are the avoidable errors? (P. Müller, Crystal Structure
Refinement, 2006)
• Wrong unit cell
• Twin refined as a disorder
• Wrong atom assignment
• Wrong space group
• Interpretation of Fourier truncation ripples as hydrogen atoms and
vice versa

And there are the “really avoidable errors” (Roland Boese, 1999):
• Typographic errors in the unit cell
• Errors in the refinement
• Wrong data collection strategy
• Data collection at room temperature

Documentation and validation

1. Documentation of an X-ray diffraction study

2. The CIF file

3. Values to judge the quality of a structure

4. Aids for structure verification

3
 Documentation of an X-ray diffraction study
As every scientific experience, an X-ray diffraction study has to be documented
according to good scientific principles.
1. Lab book: Use a dedicated lab book for the determination of structures. The lab
book should contain information not included in the electronic data (such as
crystal size, color etc.) as well as your choices made during refinement.
2. Report: The results have to include:
• Crystallisation details (solvent, temperature, how did you obtain the crystal)
• Details of the data collection (instrument, number of reflections, crystal size)
• Details of the solution/refinement (confidence factors, hydrogen treatment)
• Details of the structure (atom positions, thermal parameters, bond distances
and angles)
• A figure showing the atom numeration
• A figure showing the thermal ellipsoid (ORTEP) (could be combined)
3. Archiving: You are responsible to archive reliably (min. 5 years)
• experimental results (images etc)
• solution/refinement results (files .res, .hkl, .cif, .lst)
(A reliable archiving includes that one can find the relevant data after
several years, even in your absence!)

Documentation and validation

1. Documentation of an X-ray diffraction study

2. The CIF file

3. Values to judge the quality of a structure

4. Aids for structure verification

4
 The CIF file
CIF (Crystallographic Information File):

• S. R. Hall, F. H. Allen, I. D. Brown Acta Cryst. 1991, A47, 655-685

• Contains nearly all the information regarding an X-ray
diffraction study structure
• Nearly all journals require a CIF file as supporting information
• CIFs are used directly for publication in Acta Cryst.
• Automated validation of a diffraction study
• Automated report generation

The CIF file

data_recp1 Text file: only 80 characters per line,
only simple ASCII code.
_audit_creation_method SHELXL
_chemical_formula_moiety
“C35 H40 N6 O, C H Cl3” Each data bloc starts with data_nom.
_chemical_formula_sum Several data blocs can be combined in
'C36 H41 Cl3 N6 O' the same file.
_refine_special_details
; A variable has the format _nom.
Refinement on F^2^ for ALL
reflections except for 0 with The value of a variable can appear after
very negative F^2^
the variable separated by a space (or
;
loop_ spaces) or on the next line.
_geom_bond_atom_site_label_1 If a value contains spaces, it has to be
_geom_bond_atom_site_label_2
_geom_bond_distance included in single or double quotation
_geom_bond_site_symmetry_2 marks. If a value contains several lines
_geom_bond_publ_flag of text, it has to be separated by
N1 C1 1.346(3) . ?
N1 C5 1.349(3) . ?
semicolons.
N2 N3 1.375(3) . ?
A list of values can be initiated with the
command command loop_ . The list
_refine_diff_density_max 0.452
_refine_diff_density_min -0.616 ends with the next variable name.

5
 Generation of the CIF file
1. A first version of name.cif is generated with the command
ACTA in SHELXTL (XL)

2. You can use XCIF (or other software) to replace unknown

values in name.cif with values included in another file, such
as name.pcf (generated by XPREP). Attention: XCIF does not
update values, it only replaces unknown ones (i. e. question
marks).

3. Inspection and manual manipulation of the CIF file (using

values from other files, your lab book etc.)

4. If you add another refinement cycle, you have have to repeat

steps 2 + 3.
If you need to modify your CIF, use a simple text editor (Notepad,
Wordpad). Do not use a text editing software, such as Word! Such
software might introduce formatting codes on saving, which are not
included in the CIF format.

Meaning of variables in the CIF

What does a variable stands for? Which values are allowed?

Information can be found:

• Chapter 4.1 of International Tables for Crystallography Volume G, First
edition (2005)]
• http://www.iucr.org/iucr-top/cif/cif_core/index.html

_refine_ls_hydrogen_treatment
exemple.cif:
… Definition: Treatment of hydrogen atoms in the
_refine_ls_hydrogen_treatment ? least-squares refinement.
The data value must be one of the following:
refall refined all H-atom parameters
refxyz refined H-atom coordinates only
refU refined H-atom U's only
exemple.cif:
… noref no refinement of H-atom parameters
_refine_ls_hydrogen_treatment constr
constr H-atom parameters constrained
mixed some constrained, some independent
undef H-atom parameters not defined

6
 Documentation and validation

1. Documentation of an X-ray diffraction study

2. The CIF file

3. Values to judge the quality of a structure

• Rsigma, Rint, wR2, R1, GoF

• Residual electron density

• Correlation matrix elements

4. Aids for structure verification

The refinement – step by step

1. Reading of the .hkl file containing the reflection intensities
example.ins:
…
HKLF 4 example.hkl:
0 0 1 14.04 2.27 1
END 0 0 220748.33 212.60 1
0 0 3 3.91 4.31 1
0 0 4 8452.21 96.84 1
0 0 5 15.19 9.36 1
0 -1 -690345.18 728.10 1
In contrast to the
0 0 7 11.09 8.55 1
res/ins/cif files, the
0 0 8 2957.52 45.64 1
file *.hkl has a
precise format!
Standard
Miller indices deviation of the
Thus each space hkl intensity σ(I)
matters. But there is Intensity I batch number
no reason at all, why (see BASF in the
you should edit a SHELX manual)
…
hkl-file anyway. 0 0 0 End of file

7
 The refinement – step by step
1. Reading of the .hkl file containing the reflection intensities
exemple.lst:
h k l Fo^2 Sigma Why rejected

0 0 1 104.55 5.29 observed but should be systematically absent

2 0 1 74.15 6.04 observed but should be systematically absent
etc.
20975 Reflections read, of which 685 rejected

-10 =< h =< 10, -13 =< k =< 13, -18 =< l =< 18, Max. 2-theta = 137.96
23 Systematic absence violations

Inconsistent equivalents etc.

h k l Fo^2 Sigma(Fo^2) N Esd of mean(Fo^2)
2 0 0 127128.20 518.46 4 3194.33
0 2 0 65677.31 243.31 6 1223.07
etc.
14 Inconsistent equivalents
2847 Unique reflections, of which 0 suppressed

R(int) = 0.0275 R(sigma) = 0.0111 Friedel opposites merged

Rejected reflections: • Systematic absences

• Inconsistent equivalents
• Reflections with I < -3·σ(I)

Why use reflections with negative intensities?

• During the integration of the reflections (peaks), the intensities

are determined by I = Ireflection – Ibackground
• For weak reflections the result can be negative.

If we consider only reflections with I > 0 both

solutions 1 and 2 are equally probable.

Solution 1 Solution 2
I

0
-3σ

8
 Why use reflections with negative intensities?

• During the integration of the reflections (peaks), the intensities

are determined by I = Ireflection – Ibackground
• For weak reflections the result can be negative.
• Reflections with 0 < I < -3σ(I) contain useful information. (I = 0 is
in the margin of errors for these reflections.)
Solution 2 is more probable

Solution 1 Solution 2
I

0
-3σ

Why use reflections with negative intensities?

• During the integration of the reflections (peaks), the intensities

are determined by I = Ireflection – Ibackground
• For weak reflections the result can be negative.
• Reflections with 0 > I > -3σ(I) contain useful information. (I = 0 is
in the margin of errors for these reflections.)
• Reflections with I < -3σ(I) are physically impossible, indicate
problems in the determination and are thus omitted.
Solution 1 Solution 2
I Reflections with I<-3σ(I)
do not contain any
information, since their
value cannot be trusted.
0
-3σ

9
 3 σ: What connects the standard deviation with the
confidence interval ?
A random distribution of errors should follow a Gaussian distribution

1⎛ ( x − μ ) ⎞
2

1 − ⎜ ⎟
2⎝ σ ⎠
f ( x) = e μ: correct value
σ 2π σ: standard deviation

μ +σ

∫ f ( x ) = 0.68 = 68%
μ −σ
μ + 2σ
f(x)
∫ f ( x ) = 0.96 = 96%
μ − 2σ

μ + 3σ

(x-μ)/σ
∫ f ( x ) = 0.99 = 99%
μ − 3σ

99.7% probability that x is in the interval μ ± 3σ

a=1.234; σ=0.005: a = 1.234(5) → a = 1.234 ± 0.015

The refinement – Merging of reflections

1. Reading of the .hkl file containing the reflection intensities
Number of reflections measured
exemple.lst:
20975 Reflections read, of which 685 rejected

-10 =< h =< 10, -13 =< k =< 13, -18 =< l =< 18, Max. 2-theta = 137.96
23 Systematic absence violations

Inconsistent equivalents etc.

h k l Fo^2 Sigma(Fo^2) N Esd of mean(Fo^2)
2 0 0 127128.20 518.46 4 3194.33
0 2 0 65677.31 243.31 6 1223.07
etc.
14 Inconsistent equivalents
2847 Unique reflections, of which 0 suppressed

R(int) = 0.0275 R(sigma) = 0.0111 Friedel opposites merged

Number of independent reflections measured (not repeated, not

related by symmetry)
Governed by the command MERG in the ins-file (see SHELXTL
manual). We touch at this command only in special cases.

10
 The refinement – Error of merging
1. Reading of the .hkl file containing the reflection intensities
exemple.lst:
20975 Reflections read, of which 685 rejected

Rint …

2847 Unique reflections, of which 0 suppressed

R(int) = 0.0275 R(sigma) = 0.0111 Friedel opposites merged

Rint =
∑ F o
2
− Fo2 ( mean )
∑F o
2

• Rint: Merging error (measure of the precision/reproducibility)

• Possible error sources (high Rint value):
• Incorrect Laue group
• Bad or missing adsorption correction
• Crystal decomposition
• Twinning
• Goniometer problems (covered reflections, misalignment)

The refinement – Error of merging

1. Reading of the .hkl file containing the reflection intensities
exemple.lst:
20975 Reflections read, of which 685 rejected

2847 Unique reflections, of which 0 suppressed

R(int) = 0.0275 R(sigma) = 0.0111 Friedel opposites merged

Rsigma =
∑ o)
σ ( F 2

∑ o
F 2

• Measure of the signal-to-noise ratio

• In rough approximation, the structure confidence factor R1 cannot
be much better than Rsigma
• If Rint >> Rsigma, i. e. more than 2-3 times, there is a problem

11
 Confidence factors
Optimised value: M = ∑ w Fo − Fc( 2
)
2 2

(The lower M, the better is the agreement of our model with the experimental data.)

But: M increases with the number of reflections and with their intensity. It is thus
structure dependent, with well diffracting structures with high redundancy giving the
highest M values. We thus need a structure independent value.

Confidence factor, Residual, R-factor:

∑ w (F − F )
2 2 2
R2 wR2 = Rw ( F ) =2 o c

∑ w( F ) o
2 2

For statistical reasons, refinement against F2 gives R-factors approximately twice

as high than those for refinement against F. To facilitate comparison (and to
increase acceptance of the new method) SHELXTL calculates also the R-factor
based on F.

R1 =
∑F −F o c
R1
∑F o

Goodness of Fit
The GoF or GooF is another value which describes the
quality of our model:

GooF = S =
∑ w( F − Fc2 ) 2 o
2

N Ref. − N Par .
NRef.: number of independent reflections, NPar.: number of parameters

In contrast to the R-factor, which also depends on the signal-to-

noise ratio, S is relatively independent from the noise.

S should be around 1.
S > 1: bad model or bad data/parameter ratio
S < 1: model is better than the data: problems with the absorption
correction, space group problems

12
 Criteria for a good structure
SHELXL calculates 4 confidence values:
• wR2 (all data) Refinement against F2 requires a
• wR2 (observed data, I>2σ(I)) correct weighing scheme
• R1 (all data) The weighing schemes optimised
• R1 (observed data, I>2σ(I)) for refinement against F2 cannot
be used for the calculation of R1.

The important values are wR2 (all data) (since we do the refinement with all
data) and R1 (observed data), for comparison with the old method.

Good Acceptable Problematic Really problematic

R1 < 5% < 7% >10% >15%
wR2 < 12% < 20% >25% (ou > 2*R1) >35%
S 0.9-1.2 0.8-1.5 <0.8 ou >2 <0.6 ou >4

Residual electron density

If our model is good, we should have described all electrons in our
structure. Thus the remaining electron denisty should be zero.

Acceptable values for residual electron density:

• For light atom structures (H – F) : < 0.5 e–/Å3
• For heavy atom structures : 10% of the électrons of the heavy atom
per Å3 in a distance smaller 1.2 Å from the heavy atom. (Fourier
truncation errors)
• Accumulation of electron density on special positions

Sources of errors:
• Bad absorption correction
• Disorder

XP: info Q1 How to find the residual electron density?

test.lst:
test.cif: Electron density synthesis with coefficients Fo-Fc
_refine_diff_density_max 0.425 Highest peak 0.42 at 0.2140 0.0398 0.5135 [ 0.40 A from C2 ]
_refine_diff_density_min -0.560 Deepest hole -0.56 at 0.1424 0.4014 0.6865 [ 0.53 A from F3 ]

13
 Correlation matrix elements

test.lst:
Largest correlation matrix elements

0.853 U11 Fe1 / OSF 0.728 U11 S2 / U11 Fe1 0.524 U11 S1 / U11 Fe1
0.771 U11 S2 / OSF 0.588 U11 S1 / OSF 0.543 U12 S1 / U23 S1

Values > 0.5 for the correlation matrix elements indicate that some
parameters in the refinement are dependent from each other.
Some dependances are acceptable, such as f. e. between the
thermal parameters of the heavy atom and the overall scale factor
or between the Uxy of the same atom.
Attention: A high number of dependances > 0.5 between multiple
atoms might indicate a missed symmetry! (wrong space group)

Manual inspection : Thermal parameters

With the exception of a wrong space group, most other problems

of a structure are more visible in the thermal parameters than in
the atom positions. For bad structures, thermal parameters serve
as the garbage bin of the refinement.
In general:
• Values of the thermal displacement should be comparable for
comparable atoms.
• The displacement should be in agreement with the thermal
vibration of lowest energy.

14
 Manual inspection : Thermal parameters
Disorder
Displacement parallel to a bond greater than
perpendicular displacement
⇒ disorder, two atoms with different
bond lengths sharing the same position,
here F et Cl.
CHECKCIF: Violation of the “Hirshfeld test”

*.lst:
Principal mean square atomic displacements U
[…]
0.3098 0.0893 0.0464 C4 may be split into 0.6218 0.2673 0.2408 and 0.6118 0.2471 0.2666
0.3100 0.0924 0.0392 C5 may be split into 0.5976 0.3191 0.3424 and 0.5834 0.3017 0.3597

Manual inspection : Thermal parameters

Wrong atom
assignments
CHECKCIF: Violation of
the “Hirshfeld test”

N
N N

15
 Bond Distances

Verify if the obtained geometry is reasonable

For example, M-CH3 versus M=O:

M CH3

synthesized crystallized

Sources for bond distances and angles values:

• Cambridge data base
• A. G. Orpen, L. Brammer, F. H. Allen, O. Kennard, D. G. Watson,
R. Taylor, J. Chem. Soc. Dalton Trans. 1989, S1-S83.
• F. H. Allen, O. Kennard, D. G. Watson, L. Brammer, A. G. Orpen,
R. Taylor, J. Chem. Soc. Perkin Trans. II 1987, S1-S19.

Documentation and validation

1. Documentation of an X-ray diffraction study

2. The CIF file

3. Values to judge the quality of a structure

4. Aids for structure verification

16
 Aids to verify your structure

The CIF file enables an automated verification of your structure

• Using PLATON (http://www.cryst.chem.uu.nl/platon)
• Better, at: http://journals.iucr.org/services/cif/checkcif.html

Verifications:
• Missed symmetry (wrong space group)
• Holes/voids in the structure
• Thermal parameters (Hirshfeld test)
• Bond distances and angles
• Atom assignment
• Other details of the refinement and the data collection

There are absolutely no reasons, nor excuses,

not to verify a structure with CHECKCIF!

The Checkcif report

The automated report contains three types of alerts:

ALERT level A = In general: serious problem

ALERT level B = Potentially serious problem
ALERT level C = Check and explain

It is impossible to explain all possible CHECKCIF errors here. If you use

the checkcif routine on the site www.iucr.org, you can click on the alert to
obtain more information. Or ask other crystallographers. Or me. But
there is no excuse for ignorance.

If you ignore or comment an error, which you do not understand,

you are not an expert, but a scientific failure!

Try to eliminate all problems. If not, comment on the error, explaining

where this alert is coming from and why we can ignore it or cannot do
anything about it.

17
 Example of a checkcif report

The following ALERTS were generated.

Each ALERT has the format test-name_ALERT_alert-type_alert-level.
Click on the hyperlinks for more details of the test. Redo the refinement
with BOND $H instead
Alert level A of BOND
PLAT761_ALERT_1_A CIF Contains no X-H Bonds ...................... ?
PLAT762_ALERT_1_A CIF Contains no X-Y-H or H-Y-H Angles .......... ?

Alert level B
PLAT230_ALERT_2_B Hirshfeld Test Diff for C24 - C25 .. 7.36 su

Alert level C
CRYSC01_ALERT_1_C The word below has not been recognised as a standard
identifier. : plates
CRYSC01_ALERT_1_C No recognised colour has been given for crystal colour.
PLAT222_ALERT_3_C Large Non-Solvent H Ueq(max)/Ueq(min) ... 3.08 Ratio
PLAT230_ALERT_2_C Hirshfeld Test Diff for C25 - C26 .. 5.93 su
Correct manually
.cif: in the CIF file

_exptl_crystal_description red
_exptl_crystal_colour plates

Correcting the CIF file

It is often your job to manually correct wrong entries in

the CIF.

But: You do have to know which values you are

allowed to change and have a good reason for
changing them. Do not change the CIF or alter your
refinement just to make CHECKCIF alerts vanish!

If you are not sure: Ask!

18
 Example of a checkcif report

The following ALERTS were generated.

Each ALERT has the format test-name_ALERT_alert-type_alert-level.
Click on the hyperlinks for more details of the test.

Alert level A

Alert level B
PLAT230_ALERT_2_B Hirshfeld Test Diff for C24 - C25 .. 7.36 su

Alert level C
PLAT222_ALERT_3_C Large Non-Solvent H Ueq(max)/Ueq(min) ... 3.08 Ratio
PLAT230_ALERT_2_C Hirshfeld Test Diff for C25 - C26 .. 5.93 su

Short detour: Hirshfeld rigid-bond test

Anthony L. Spek (author of PLATON), Acta Cryst. (2009). D65, 148–155
“The Hirshfeld rigid-bond test (Hirshfeld, 1976) has proved to be very
effective in revealing problems in a structure. It is assumed in this test
that two bonded atoms vibrate along the bond with approximately equal
amplitude. Significant differences, i.e. those which deviate by more
than a few standard uncertainties from zero, need close
examination. Notorious exceptions are metal-to-carbonyl bonds, which
generally show much larger differences (Braga & Koetzle, 1988).”
Hirshfeld, F. L. (1976). Acta Cryst. A32, 239–244.
Braga, D. & Koetzle, T. F. (1988). Acta Cryst. B44, 151–156.

19
 Hirshfeld test violations
Wrong atom assignment:
Bad description of thermal ellipsoids
(data quality problem, disorder etc.)

Substitutional disorder:

M M
but:
Vibration of a
but: M whole group

M=C=O
Linear distribution of electrons in bonds
Braga, D. & Koetzle, T. F. (1988). Acta Cryst. B44, 151–156.

Example of a checkcif report

Alert level B
PLAT230_ALERT_2_B Hirshfeld Test Diff for C24 - C25 .. 7.36 su
=> caused by vibration of a complete phenyl substituent

Alert level C
PLAT222_ALERT_3_C Large Non-Solvent H Ueq(max)/Ueq(min) ... 3.08 Ratio
=> presence of strongly vibrating terminal phenyl groups and strongly bound Cp-
groups

PLAT230_ALERT_2_C Hirshfeld Test Diff for C25 - C26 .. 5.93 su

=> caused by vibration of a complete phenyl substituent

C24

C25

C26

20
 The Checkcif report
Why should you comment on CHECKCIF errors:
• You have to do it anyway, if you want to publish a structure.
• It shows the reader that, despite the errors, the structure is of good
quality.
• When you write up your thesis or a publication, sometimes years
later, you might not recall what caused the alert and what you have
already tried to resolve it. Even worse, if you are gone and your
supervisor tries to publish your results.
• Finish all structures to the point, where they are immediately
publishable.

How to comment on remaining CHECKCIF errors:

1. Copy the report in a Word or equivalent software, add your comments
and archive it together with the CIF. Not the best choice, since the
information is not in the CIF.

The Checkcif report

How to comment on remaining CHECKCIF errors:
2. Write your comments in the CIF file. For example, under the
variable _refine_special_details :
_refine_special_details
;
Refinement of F^2^ against ALL reflections. The weighted R-factor wR and
goodness of fit S are based on F^2^, conventional R-factors R are based
on F, with F set to zero for negative F^2^. The threshold expression of
F^2^ > 2sigma(F^2^) is used only for calculating R-factors(gt) etc. and is
not relevant to the choice of reflections for refinement. R-factors based
on F^2^ are statistically about twice as large as those based on F, and R-
factors based on ALL data will be even larger.

Comments on remaining CHECKCIF errors:

Hirshfield test violations for atoms C24, C25 and C26 are caused by a
strong vibration of a complete phenyl substituent.

The large difference between Ueq(max)/Ueq(min) is caused by the presence

of strongly vibrating phenyl groups and strongly bound Cp atoms in the
structure.
;

Advantage: already present in CIF. Disadvantage: Easily missed by reviewers.

21
 The Checkcif report
How to comment on remaining CHECKCIF errors:
3. Add your comments using the “Virtual reply form” :
Checkcif report:

Alert level B
PLAT230_ALERT_2_B Hirshfeld Test Diff for C24 - C25 .. 7.36 su
Alert level C
PLAT222_ALERT_3_C Large Non-Solvent H Ueq(max)/Ueq(min) ... 3.08 Ratio

CIF file:
data_example1

_vrf_PLAT230_example1
;
RESPONSE:
Caused by vibration of a complete phenyl substituent
;

_vrf_PLAT222_example1
;
RESPONSE:
Presence of strongly vibrating terminal phenyl groups and
strongly bound Cp-groups.
;

The Checkcif report

Disadvantage: A bit more effort, the format must be followed exactly!
_vrf_error code_data code
;
RESPONSE:
Your reply
Your reply
Your reply
;

Datablock: shap30

The following ALERTS were generated. Each ALERT has the format test-name_ALERT_alert-type_alert-
level. Click on the hyperlinks for more details of the test.

Alert level B PLAT230_ALERT_2_B Hirshfeld Test Diff for C24 - C25 .. 7.36 su
Author Response: Caused by vibration of a complete phenyl substituent

Alert level C PLAT222_ALERT_3_C Large Non-Solvent H Ueq(max)/Ueq(min) ... 3.08 Ratio

Author Response: Presence of strongly vibrating terminal phenyl groups and strongly bound
Cp-groups.

Advantage: Your comments will be visible if anybody submits

the CIF file to the CHECKCIF routine

22
PLAT154_ALERT_1_G The su's on the Cell Angles are Equal .......... 0.00200 Deg.
Author’s response: Provided data is correct. Unrounded su’s show slight differences.
CIF:
_cell_angle_alpha 92.702(2)
_cell_angle_beta 93.216(2)
_cell_angle_gamma 98.523(2)
A B C Alpha Beta Gamma Vol
9.40774 9.44160 21.46292 92.7024 93.2165 98.5227 1879.43
0.00006 0.00006 0.00015 0.0002 0.0003 0.0002 0.03
Corrected for goodness of fit:
0.00057 0.00058 0.00135 0.0023 0.0025 0.0021 0.23

PLAT761_ALERT_1_A CIF Contains no X-H Bonds ...................... ?

PLAT762_ALERT_1_A CIF Contains no X-Y-H or H-Y-H Angles .......... ?

Add a BOND $H command to the INS file. Repeat the refinement.

PLAT093_ALERT_1_B No su's on H-atoms, but refinement reported as . Mixed

Enter the correct value, most likely constr

PLAT250_ALERT_2_C Large U3/U1 Ratio for Average U(i,j) Tensor .... 3.5
Author Response: Manual inspection of the thermal ellipsoids does not indicate any
preferred orientation. The increased U3/U1 ratio is most probably caused by the
toluene solvent disordered around the inversion center, which has increased thermal
ellipsoids parallel to the line connecting the centers of gravity of the disordered
molecules.

23
 PLAT232_ALERT_2_C Hirshfeld Test Diff (M-X) Cr1 -- N1 .. 6.0 su
Author Response: Manual inspection of the thermal ellipsoids does
not indicate any problems. The alert might be caused by the low sd of
the heavy atom position.

Since the Hirshfeld test is performed relative to the esd of the thermal parameters,
heavy atoms with very low esds show sometimes “invisible” Hirshfeld test
violations.

PLAT072_ALERT_2 SHELXL First Parameter in WGHT Unusually Large. 0.12

Author Response: PLATON TWINROTMAT did not indicate any evidence for twinning.

PLAT366_ALERT_2_C Short? C(sp?)-C(sp?) Bond C26 - C27 ... 1.38 Ang.

Author Response: Styrene olefinic bond. For discussion, see text.

PLAT241_ALERT_2_A Check High Ueq as Compared to Neighbors for C69

Author Response: We observed deviations of expected thermal parameters which are in
agreement with a slight rotational disorder of the benzyl groups. (ADPs decrease from para
to ipso). The disorder is present in all benzyl groups and affects the molecule as a whole.
Resolution of the disorder was possible using very strong geometric restraints, but even
upon application of strong restraints on the thermal parameters, anisotropic refinement of
the solved disorder was not possible. The isotropic, disordered model did not improve the
structural quality (as judged from variances of chemically equivalent bond lengths) and the
structure was refined anisotropically with the disorder unresolved to provide the
uncertainities of atom positions in ORTEP plots and bond precision. No indication of
twinning was observed during integration or in post-refinement checks (TWINROTMAT etc.)

C69

24
 Checklist
‰ wR2 < 12%
‰ R1 < 5%
‰ Rint < 3 x Rsigma
‰ 0.9 < GoF < 1.2
‰ Residual electron density < 0.5 e-/Å3
‰ No correlation matrix elements > 0.5
‰ Thermal ellipsoids are acceptable
‰ Checkcif, checkcif report and explications
‰ Figure showing thermal ellipsoids (50% probability level) and
the numeration of atoms
‰ Printed version of the structural details, bond distances and
angles
‰ Archiving: 1 x with the X-ray service, 1 x for you, 1 x for the
synthetic chemist, 1 x for your supervisor

The end

25