Hi Pavel,

can't give you an expert opinion, because never dealt with TEM diffraction, but I guess it doesn't differ too much from any other diffraction in terms of physics, does it? - apart from the fact that (as far as I know) TEM works mostly in kinematical approximation for electron-matter interaction, where the dynamical effects are negligible, that's all due to the high energy of the electron beam and small thickness of the sample. Therefore, I would definitely expect the dynamical effects, if any, to be quite moderate, meaning that whatever reflections originate from those, their intensity would be several orders of magnitude lower, than that of primary spots. In the pattern you provide, the superstructure spots are way too bright, in my opinion, to originate from some tiny effects like "forbidden reflections", etc... So, from the general knowledge position, I'd suggest you optimize your model. Might that be you have some surface reconstruction? Also, the indexing of the EDP_experiment and .._theory is a bit confusing - but I guess the indices are supposed to be the same/similar?

Hope this helps to some extent

Best regadrs

Nikolay

Hi Pavel,

In your diffraction patterns, the experimental patterns seem to be taken from [010] zone axis and theoretical is shown in [100]. Unless your material is cubic, how can you compare both patterns?

Anyway for understanding the superstructural or superlattice reflections better, may be you can perform diffraction tilting experiment and observe the changes in the reflections in different zone axes. Based on that it is possible to construct the unit cell and understand the structure better.

Dear Pavel,

You are right: double diffraction cannot explain the superstructure reflections (see later on). Also explanation by dynamical effects can be discarded.

Let me briefly discuss the two SAD patterns: The setting of the experimental pattern differs from that of the theoretical pattern:

a(exp) =-c(theo),

b(exp) = a(theo)(=0=-a(theo), to keep the determinant positive)

c(exp) = b(theo)

In the following I will refer to the indexing of the t_h_e_o_r_e_t_i_c_a_l pattern, i.e. the zone is [1 0 0].

To discuss your problem it could be helpful to know something about your model, in particular about the 3rd dimension and the space group(SG). I suppose the structure is orthorhombic. Anyhow, an explanation of the "forbidden reflections" by dynamical effects can be discarded.

Obviously your model implies a 21-axis parallel b or a b-glide plane * * b (hk0 present for k=2n only). The violation of this condition is the only clear indication for double diffraction (when speaking about presence and absence of reflections, not about intensisies).

An interpretation of the superstructural reflections as traces or tails of 1st order Laue-zone (FOLZ)-reflections of a B-centered structure with large cell parameter a (theor. setting) is extremely improbable.

In each case you should modify the structural model.

The following symmetry elements may lead to extinctions in the pattern of your model which are violated in the experimental SAD:

hkl B-centering, hkl only for h+l=2n, SG: B * * *

or 0kl c-glide-plane: 0kl only for l=2n, SG: P c * *

Note: B-centering cannot be violated by double diffraction!

A concluding question:

The term "superstructural reflections (1/2 1/2 0)" is somewhat confusing. Is the interpretation "replace all indices h by h/2 and k by k/2" (referring to the experimental pattern)correct? Then cell parameter a to which the new indices refer is halved. While the main reflections remain integers the superstructural reflections h=(2n+1)/2 are not integers anymore. Since k=0, k/2 is 0 as well. However, the notation (1/2 1/2 0)  suggests that b is halved as well.

Perhaps these notes may help you to revise your model.

Good luck!

With best wishes,

Gerhard Miehe