2.4.2. X-Ray Fluorescence (XRF) Analysis

To determine metal concentration (% *w*/*w*) in the composites, a VRA 30 X-Ray fluorescent analyzer (Leipzig, Germany) was used. To excite XF, an X-Ray tube with a Mo anode was used at the acceleration of 50 kV and current of 20 µA. For analysis, Cu-chitosan powders or gels in an amount of 10–12 mg

was thoroughly ground and pressed into pills. Then the XRF spectra of the composites and reference samples were recorded. The standard buffer solution for spectrometer calibration was composed of a mixture of polysterene/metal salt. Quantity analysis was conducted through comparison of the peak intensity of Cu Kα line in XRF spectrum of the composite with the values of the calibration curve obtained previously.

### 2.4.3. Conventional Small-Angle X-Ray Scattering (SAXS) Analysis

SAXS measurements were done on laboratory diffractometer "AMUR-K" (developed in A. V. Shubnikov Institute of Crystallography, Moscow [21]). Wavelength of X-rays λ = 0.1542 nm was used, applying Kratky type geometry covered the range of scattering vector modulus 0.12 < *s* < 6.0 nm−<sup>1</sup> (*s* = 4π*sin*θ/λ; 2θ is the scattering angle). Experimental data were normalized to the intensity of the incident beam, and then a correction on collimation error was made according to standard procedure [22]. Further data processing and interpretation was done using the program suit ATSAS [23].

Volume size distribution functions *DV(R)* of heterogeneities in the specimens and distance distribution functions *p(r)* were computed by means of the regularization technique realized in program GNOM [24]. The low-resolution shapes of the Cu nanoparticles in the Cu-carrying chitosan were reconstructed ab initio using distance distribution function *p(r)* and program DAMMIN [25]. The program utilizes a simulated annealing algorithm to build models fitting the experimental data *Iexp(s)* that minimizes the discrepancy:

$$\chi^2 = \frac{1}{N-1} \sum\_{j} \left[ \frac{I\_{\exp}(s\_j) - cI\_{calc}(s\_j)}{\sigma(s\_j)} \right]^2$$

where *N* is the number of experimental points, *c* is a scaling factor and *Icalc(s<sup>j</sup> )* and σ*(s<sup>j</sup> )* are the calculated intensity from the model and the experimental error on the intensities at the momentum transfer *s<sup>j</sup>* , respectively. The program DAMMIN was run of about a dozen separate calculations to identify the most typical models.
