From the data acquisition process, the following data will have been obtained:

• Photographs of the crystal

• Individual diffraction frames

• Unit cell parameters

• A HKL file

• The structure output from AutoChem

• Information obtained during the pre-experiment acqusition

A crystallographer will then carry out the structure solving steps. The crystallographer is effectively creating a model which would result in the observed diffraction pattern from the data collection.

The individual diffraction frames are analysed by computer to combine the frames into a single file to use within the processing. The data is combined into a .run file which details each individual frame, diffraction pattern locations and intensities.

As part of this process 'Peak Hunting' works through each frame and recording the unit cell matching. Whilst not every frame will give perfect agreement with the unit cell, the overall level of unit cell matching should be high.

Using the run fine, the individual frames are then combined into an electron density map showing the intensities and positions of electron density. At this stage, the phase of the wave has not been determined. Needs more explanation, including how this is solved...

Twinning would be considered at this stage by combining offset data from the twins (two or more crystals).

In most cases, the detected options are usually taken initially, but this step could be revisited later in the process if changes need to be made.

Individual frames could also be removed at this stage, potentially due to bad icing, but usually this would be carried out later, during the structure refinement.

Datasets from multiple crystals of the same compound could also be combined in this stage where necessary.

The software identifies hkl1, and the crystallographer manually assigns the crystal dimensions from the photograph and hkl intersects. This gives the dimensions of the crystal, which is then used in the crystal modelling. This process is repeated for many rotations of the crystal, until the computer is happy that the dimensions have been determined for the crystal.

At this stage a RES file (INS file) is generated which contains all the hkl and structure factor data, with these combined into a single file which can be used to solve an initial model.

This step is usually carried out in OLEX2. The solftware uses various models to solve the structure, which includes the phase solving. Generally ShelXT is used initially, but other models can be used if nescessary.

From the initial solving attempt using a computational model in the step above, the crystallographer can then consider how to refine the model, applying human expertise in the process. There are a number of steps the crystallographer may work through, including trying alternative models from the step above. The steps below are likely to be carried out roughly in order, but steps may be revisited in the process. Earlier steps may also be revisited, such as the removal of additional data frames where nescessary.

Throughout the images below, the yellow-orange spheres indicate residual electron density which has not been accounted for in the model. These are referred to as Q-peaks in crystallography. As the model is gradually improved, fewer Q-peaks remain shown.

The crystallographer can use their chemical knowledge to alter the automatic atom assignment. This involves both information provided through the sample submission form providing information on the expected structure, and through application of other chemical knowledge. In the step above, note that the left hand ring has what appears to be a carbon atom with two oxygens attached, which is actually a nitro group. In the image below, the crystallographer has manually corrected the atom assignment for this. Nitrogen and carbon only differ by one electron, so this is a fairly common and understandable misassignment, from the automated models.

Initially each atom is isotropic and is modelled with minimal parameters. The use of orbital correction adds additional parameters which better reflect reality by plotting the atoms aniostropically, which takes into the thermal parameters of each atom.

In the model, not every molecule in the crsyal structure sits in exactly the same alignment. This compensates for small amounts of disorder in the individual molecules in the structure, for example rotation of atoms (eg CF3 groups) or disorder in hexyl or isopropyl chains etc

At this stage it is usual to assign hydrogen atoms onto the structure, which will see more of the residual electron desity (Q peaks) being replaced by the H atoms.

Where residual density (Q peaks), located far from the structure, the model is refined to try assign the solvent.

Could be multiple molecules. These could be mirror images, rotations, improper rotations, glides which are symmetry operations and determined by the space group. These could also be non symmetry operations, where more than one molecule of the compound exists in the unit cell.

including unit cell parameters, angles and volume 

Assigns atom labels and numbers, and this is refined to make sense. Eg numbering related structures in the same manner.

Meta data added.

Crystallographer checks stuff makes sense. Return to user

Crystal structures are routinely deposited into crystal structure databases which allow other researches to access and use the data obtained from single crystal X-ray diffraction. 

Information on using CIF files and obtaining information from crystal structures can be found in the X-ray data pages.