Supercritical Antisolvent Processing of Nitrocellulose: Downscaling to Nanosize, Reducing Friction Sensitivity and Introducing Burning Rate Catalyst

A supercritical antisolvent process has been applied to obtain the nitrocellulose nanoparticles with an average size of 190 nm from the nitrocellulose fibers of 20 μm in diameter. Compared to the micron-sized powder, nano-nitrocellulose is characterized with a slightly lower decomposition onset, however, the friction sensitivity has been improved substantially along with the burning rate increasing from 3.8 to 4.7 mm·s−1 at 2 MPa. Also, the proposed approach allows the production of stable nitrocellulose composites. Thus, the addition of 1 wt.% carbon nanotubes further improves the sensitivity of the nano-nitrocellulose up to the friction-insensitive level. Moreover, the simultaneous introduction of carbon nanotubes and nanosized iron oxide catalyzes the combustion process evidenced by a high-speed filming and resulting in the 20% burning rate increasing at 12 MPa. The presented approach to the processing of energetic nanomaterials based on the supercritical fluid technology opens the way to the production of nitrocellulose-based nanopowders with improved performance.


S3
Section S1. Experimental Methods Electron microscopy. Target-oriented approach was utilized for the optimization of the analytic measurements [1]. Before measurements the samples were mounted on a 25 mm aluminum specimen stub and fixed by conductive adhesive. Metal coating with a thin film (10 nm) of platinum/palladium alloy (80/20) was performed using magnetron sputtering method as described earlier [2]. The observations were carried out using Hitachi SU8000 field-emission scanning electron microscope (FE-SEM). Images were acquired in secondary electron mode at 2 kV accelerating voltage. Morphology of the samples was studied taking into account possible influence of metal coating on the surface [2].
Samples morphology was studied also using Hitachi transmission electron microscope (TEM).
Before measurements the samples were mounted on a 3 mm copper grid and fixed in a grid holder.
Images were acquired in bright-field TEM mode at 100 kV accelerating voltage.
Atomic force microscopy. Morphology of the samples was revealed by scanning probe microscopy (NTEGRA Prima microscope, NT-MDT) with ≤ 10 nm and ~1nm cantilevers. Powdered samples were pressed in small pellets and its surface was monitored in semi-contact or contact scanning modes.
Specific surface area. The BET surface area was determined using FlowSorb III 2305 (Micromeritics) instrument by measuring the adsorption of a gas mixture (30%N2/70%He) on the surface of powder. Friction sensitivity was determined in full set of experiments according to STANAG 4487 [3].

Section S2. Starting Materials
Scanning electron and probe microscopy have been used to disclose the morphology of the involved powders. Figures S1a, b represent the view of as-received nitrocellulose fibers showing a smooth surface with some cracks. Carbon nanotubes appeared as the elongated tortuous objects. In line with the manufacturer specification [4] the inner diameter of CNTs is ca.10 nm, the outer diameter is near 20 nm, and the length is no less than 2 μm (Figures S1c, b). Iron oxide is formed mainly by nanoparticles as evidenced from both types of microscopy ( Figure S1e

S6
Section S3. Details of the SAS process Table S1 summarizes the performed experiments with identification of the following governing parameters: the nozzle diameter D, autoclave pressure P, the ratio between solution (suspension) and CO2 feeding rates (f/F), the volume of solvent V, the NC mass m(NC) and its concentration in solution (suspension), w(NC), the additive type present, and the characteristics of the final product -its morphology and yield (N stands for number of runs). Nickel and copper salicylates were intended to be introduced same as the iron oxide since they are known as catalysts for NC-based propellants [5]. But both salts appeared to be soluble in acetone-CO2 binary system, thus not suitable for the selected process.

Section S4. Morphology of Nano-NC
Particle size distribution for nano-NC powder has been obtained by counting particles in AFM images. Typical AFM image is presented in Figure S2 and details of the numerical particle-size distribution can be found in Table S2.

Section S5. Determination of catalyst content in fabricated composites
Considering the non-100% recovery of the product for NC composites with additive (see Table S1), the exact additive content in material can differ from that in original suspension, i.e., 1 wt.% of CNTs and 5 wt.% of nano-Fe2O3 above the 100% of NC. To check if this content remains the same in the final product, we perform the gentle annealing of the composites to receive the iron oxide residue. Figure S3 illustrates the heating program, the key issue is to use the low heating rates and sample masses during NC-core decomposition to prevent ignition of the sample and low recovery of the residue (e.g., 0.5 wt.% after 10 K/min heating to 300°C in preliminary tests). Sample loads were near 10 mg, low enough to eliminate self-heating, and appropriate to give a sufficient accuracy of weighting the residue.

S10
Section S6. Thermokinetic analysis Figure S4 shows the results of formal kinetic analysis of DSC curves obtained at 0.5-5 K/min for micron-(a) and nano-(b) NC. Non-linear regression estimates of the kinetic parameters and information criteria showing that single-step kinetic models outperform the two-step models are presented in figures.  Figure S4. Results of formal kinetic analysis for micron-(a) and nano-(b) sized nitrocellulose.
The obtained kinetic parameters were compared with the extensive data from review by Brill and Gongwer [7]. Figure S5 shows that our results fall to the kinetic compensation line that is supposed to include the autocatalytic reaction parameters. Present study Figure S5. Data from review [7] and the results of the current study.

S12
Section S7. Combustion tests Figure S6 shows the images of pressed nitrocellulose samples before and after experiment, as well as a frame from video during combustion. For nano-sized NC the amount of residue is notably higher that for micro-sized sample. Analogously, the views of the samples before, during, and after combustion are presented in Figure   S7. The additive type apparently influences the appearance of the combustion surface, the view and amount of the combustion residue.