Aller au menu principal Aller au contenu principal

SWING, Small and Wide angle X-ray scattering

For studying complex structure samples: macomolecules, nanomaterials, tissues...


Website under construction... please be patient

By providing information on the structure of matter at scales varying between nanometer and micrometer, the beamline SWING will help answering the numerous questions related to soft condensed matter, conformation of macro-molecules in solution and composite materials in material sciences.

This experimental set up will allow small-angle X-ray scattering (SAXS) and wide-angle X-ray scattering measurements (WAXS) to be performed simultaneously in the 5-17 keV energy range, as well as grazing incidence small angle scattering (GISAXS). Anomalous scattering experiments will be facilitated. Emphasis will be put on the variety of samples that can be studied, solutions, gels, amorphous solids, crystallised solids and the corresponding diversity of sample environments.


Post Doctoral Student
Beamline Scientist
DOURI Nadine
Beamline Engineer Assistant
PEREZ Javier
Beamline Manager
THUREAU Aurelien
Beamline Scientist

Technical data

Energy range

Between ~5 and 17 keV

Energy Resolution

~2 eV


In-vacuum U20 undulator.
Source Size (sigma, μm): 388 (H) x 8.1 (V)
Source Divergence (sigma, μrad): 14.5 (H) x 4.6 (V)


Diaphragm at 11.7 m (1x0.5 mm2
Fixed exit DCM Si111at 20 m
Fixed incidence focusing KB at 22.5 m 
Sample position : 30 - 32 m
Detector / Sample Distance : 0.5 – 8 m

Sample Environments

X / Z precision table
Stopped flow device for chemistry
Online HPLC for proteins in solution (SAXIER project 

High throughput sampler for proteins in solution (SAXIER project)
Couette Cell (collaboration with LPS, Orsay)
GISAXS chamber 
Automatic sample changer (50 samples, thermostated)
Linkam heating stage THMS600

Beam size at sample

450x20 μm2 FWHM in the experimental hutch

Flux on sample

8.1012ph/s @7keV, 8.1011 ph/s @16keV (with 400 mA ring)


SAXS : PCCD170170 (AVIEX), Gain > 3ADU/ph, Noise≈2ADU 
WAXS (2014) : Hybrid Pixel detector (Detector Group pool)

Detection chamber

Under primary vacuum, SAXS detector positions :
- 0.20 / + 0. 20 m (horiz), -0.20 / +0.20 (vert), 0.5 m / +6 m (along X-rays).

Scientific opportunities

To be filled


BioSAXS Course - Javier Pérez :  

Foxtrot Software

Foxtrot Software (Data reduction and treatment)

Ask by mail

Foxtrot tutorial :
Foxtrot : F.A.Q.

Denfert Software

Denfert Software (Ab initio shape determination, including hydration layer)
version 2.0 (June 2015)

Denfert : F.A.Q.
A zip archive containing executables for Mac, Linux and Windows together with a test example may be downloaded by clicking here.

... a program for the ab initio “dummy-atom” structural modeling of Biological Macromolecules including the contribution of their inherent hydration layer.

Developed by Alexandros Koutsioubas and Javier Pérez

If you use DENFERT, please cite :

A. Koutsioubas & J. Pérez, Journal of Applied Crystallography (2013) 46, 1884 and A. Koutsioubas, S. Jaksch & J.Pérez, J. (2016). J. Appl. Cryst. 49  (doi:10.1107/S1600576716003393)

Brief Description

DENFERT is implementing a simulated annealing algorithm similar to DAMMIN program by D. Svergun (Biophys. J. 76, 2879-2886) for the restoration of low-resolution structural info of bio-molecules from SAXS and SANS data. The major advantage of DENFERT is that the hydration layer around bio-molecules is taken into account by introducing a second type of beads (hydration beads) in the model.

In the top figure, we see an example of the shape restoration of Lysozyme from SAXS data using DENFERT. A cartoon representation of the crystallographic structure is also presented for comparison. The bottom figure depicts additionally the hydration layer around the reconstructed protein shape.

User Manual

In order to run the program, the user should provide the GNOM (ATSAS suite) output file, and the electron or scattering length densities of the molecule, hydration layer and solvent in e/A^3 or 10^-10 cm^-2. Scattering data should be provided in (A^-1) units. During the run and at each temperature step, the program writes on the disk the current system configuration in two pdb files (model, hydration layer), and a log text file. After the program has finished an additional ASCII file containing the experimental data and the fit is written on the disk. All output filenames begin with the project name provided by the user. Output pdb files can be visualized with standard programs like pymol and may also serve as input to the analysis tools provided by the ATSAS suite.

By default the program runs in Dialog mode. However a parameter ASCII file can be provided as command line argument, containing user input in exactly the same order as asked in the Dialog mode. Each parameter should be placed in a newline in the input file.

During input, the user is asked to specify a loosness penalty weight that is related to the final compactness of the structure. The default value is a safe choice, but the parameter can be relaxed at will.

Running the program in Fast/Slow or Expert mode will affect the size of the beads and concequently their overall number during the annealing procedure. If the annealing takes too long, consider to increase the size of the beads.

The parameter 'knots' defines the number of the points of the curve that are fitted during the simulated annealing procedure.

The annealing schedule parameter affects the speed of the temperature decrease at each annealing step (Temp=anneal. schedule x Previous Temp).

Trials per bead at each temperature is set by default equal to 100. Larger values can be set in advanced mode.

In a successful run the goodness of fit Rf should be less than 10^-1 and loosness (loose) should be less than 0.02.

Note that if qmax is larger than about 0.2-0.25 A^-1 then by default the program attempts to subtract an appropriate small constant from the experimental data in order to force q^-4 Porod behavior at higher q. In order to compare results with the program DAMMIN, the same constant may be subtracted for an identical considered q-range.

By setting the electron density of the hydration layer equal to zero, the program runs without hydration beads.

Example Program Dialog 

 *** Ab-Initio low-resolution shape determination of         ***
 *** hydrated biological molecules from SAXS/SANS data       ***
 *** DENFERT version 2.0.0, June 2015                        ***
 *** Alexandros Koutsioubas (a) & Javier Perez (b)           ***
 *** (a) JCNS outstation at MLZ, Forschungszentrum Jülich    ***
 *** (b) Beamline SWING, Synchrotron SOLEIL                  ***
 *** please reference: J. Appl. Cryst. (2013) 46, 1884       ***

run with command line argument -help for brief usage instructions


[F] fast, [S] slow, [A] advanced mode?................ A            
------ Choose run mode. Speed of executation is affected by the number of dummy atoms.
[X] SAXS data or [N] SANS data?....................... X
------ Choose x-ray or neutron scattering mode.
Project name.......................................... Lysozyme     
------ All output files will begin with the given project name
GNOM output filename.................................. gnolyz.out   
------ output filename (with extension) of the program GNOM (ATSAS suite)
Qmax (default=0.25A^1)................................ 0.25         
------ Maximum wave vector to be considered by the algorithm
Maximum Diameter - Dmax =   50.0 A ................... 50           
------ Maximum diameter of the molecule (Default value taken from GNOM file)
Particle's electron density (default=0.44e/A^3)........0.44         
------ Particle electron density (default value for protein molecules)
Solvent/buffer electron density (default=0.334e/A^3).. 0.334        
------ Buffer electron density (default value for pure water)
Hydration layer contrast (default=0.03e/A^3).......... 0.03         
------ Hydration layer contrast (default values ~10% higher than for bulk water)
Calculating contribution of internal inhomogeneities... Please wait...
Constant to subtract from SAXS data = 0.359E-01 ...... 0.359E-01         
------ Constant Subtraction in order to obtain the equivalent "shape curve"
Number of knots (default= 20)......................... 20           
------ Number of number of the points of the curve that are fitted during the simulatedannealing procedure.
Dummy atom radius (default=1.795A)....................1.795         
------ Packing radius of dummy atoms (should be kept < 4.5A if possible)
Initial Annealing Temperature (default=0.001)......... 0.001        
------ Temperature of the first annealing step
Annealing Schedule (default=0.90)..................... 0.90         
------ at each step T is the equal to the previous T times the schedule factor
Penalty weight (default=0.600E-02)................0.006             
------ Penalty weight for particle’s compactness
Trials per bead at each temperature (default=100) .... 100          
------ Max trials per annealing step (increase for even better convergence)

   Generating initial model...
 Parameters of Simulated Annealing run
 Project name : Lysozyme
 Gnom SAXS Input file : gnolyz.out
 Solvent/buffer electron density (e/A^3) :       0.334
 Hydration layer contract (e/A^3) :              0.030
 Particle electron density (e/A^3) :             0.440
 Qmax (A^-1) :                                   0.250
 Maximum Diameter of the particle (A) :           50.0
 Bead packing radius (A) :                       1.795
 Number of experimental points :                    90
 Number of knots :                                  20
 Subtracted Constant (SAXS):                 0.359E-01
 Initial Annealing Temperature :             0.100E-03
 Annealing Schedule :                            0.950
 Trials per bead at each temperature step :        100
 looseness penalty :                         0.600E-02

------ At the start of each run, a summary of parameter values is displayed

   Total number of beads of the initial model: 1505
 Initial scattering curve calculation...
 Rf^2 of the initial configuration: 0.108E+00
 Start of simulated annealing procedure:
 Initial annealing temperature: 0.100E-03
 Temp=0.100E-03 | Rf^2 =0.775E-01| Rf^2 + penalties =0.778E-01 | loose=0.407E-01
 success=   296/   714 | #beads=  713(   702) | Rg=  15.56A | Volume =  23342A^3
 acceptance ratio high... skipping to lower annealing temperature...

------ At each annealing step the model’s parameters related the score function are diplayed together with the particle’s volume and radius of gyration. Also the number of succesfull bead reconfigurations is given.

----- At the end of the annealing process, chi against the experimental curve is given and a file containing the final fit is generated.

Example of running the program in non-dialog mode

In order to run the simulated annealing session that is presented above without passing through the dialog procedure, a parameter file should be passed to the program as argument.

./denfert_linux64 <parameter filename>

the parameter file should have the following syntax





an empty line in the input file means that the default value should be used.