## Magnitude of the effective form factor (| ar{f}_H |) and its two different standard deviations (ar{V}_{
m config}^{1/2} and V_{
m time}^{1/2}) of Fe as a function of the fluence, (a) at a photon energy of 6.1 keV (below *K*-edge) and (b) at a photon energy of 8.1 keV (above *K*-edge)

**Figure 1.** Magnitude of the effective form factor (| \bar{f}_H |) and its two different standard deviations (\bar{V}_{\rm config}^{1/2} and V_{\rm time}^{1/2}) of Fe as a function of the fluence, (a) at a photon energy of 6.1 keV (below *K*-edge) and (b) at a photon energy of 8.1 keV (above *K*-edge).

**Abstract**

The high-intensity version of multiwavelength anomalous diffraction (MAD) has a potential for solving the phase problem in femtosecond crystallography with x-ray free-electron lasers (XFELs). For MAD phasing, it is required to calculate or measure the MAD coefficients involved in the key equation, which depend on XFEL pulse parameters. In this work, we revisit the generalized Karle–Hendrickson equation to clarify the importance of configurational fluctuations of heavy atoms induced by intense x-ray pulses, and investigate the high-intensity cases of transmission and fluorescence measurements of samples containing heavy atoms. Based on transmission/fluorescence and diffraction experiments with crystalline samples of known structures, we propose an experimental procedure to determine all MAD coefficients at high x-ray intensity, which can be used in *ab initio* phasing for unknown structures.