*3.1. MW Radiation Effect in the Production of NH4CN Polymers*

In order to explore the effect of the MW radiation in the aqueous polymerization of NH4CN, two set of experiments were carried out. The first one used equivalent reaction times (polymers **1–5**, Table 1); the second one fixed the reaction time at 67 min (polymers **2**, **6–9**, Table 1). For the first series of syntheses (series 1), the reaction times were chosen with a comparative proposal in a relationship with the NH4CN polymer obtained at 80 ◦C using conventional heating and a reaction time of 144 h (control polymer) [1]. Based on the manufacturer specification of the MW reactor, 6 days under a classical thermal treatment is equivalent to the short reaction times using MW radiation indicated in Table 1 (series 1). On the other hand, the fixed time for the second series (series 2) was based on the fact that, in order to probe a polymerization time shorter than 309 min (polymer **1**) but longer than those explored previously in the NH4CN polymerization assisted by MW at 180 ◦C, with a reaction time of up to 30 min [13]. Thus, the reaction time for polymer **2** was chosen for polymeric series 2.

The conversion degrees, α (%), reached for each polymer are displayed in Figure 2a, and they were calculated as is described in [1] by means of gravimetric measurements of the gel fraction and the insoluble product. The use of conventional heating demonstrated that the increase in the temperature leads to a decrease in the yield of the NH4CN polymers [14]. For the control polymer, the conversion was about ≈38%, while this value decreased to ≈17% when the polymerizations are conducted at 180 ◦C using MW radiation [13]. Note that the higher conversion was found in series 1, around 25%, at the lowest temperature under study 130 ◦C, and that the data of series 1 is in agreement with these previous results. However, a slight increase in the conversion for the polymerizations was observed at temperatures higher than 180 ◦C. For series 2, at 130 ◦C, a decrease of practically five times in the polymerization time produced a relevant and unexpected increase in conversion. On the contrary, an increase in the reaction time at temperatures of 170 and 205 ◦C causes an increase in the yield of insoluble polymer. The conversion values for polymers **6**, **7** and **9** was really unexpected, reaching yields similar to those obtained by means of a traditional heating process [1,14]. This behavior of the polymerization in the presence of MW radiation seems to indicate significant differences with respect to the kinetic approach reported using conventional heating, in which a clear limit conversion was observed after a concrete reaction time for a fixed temperature [4]. Further works are in progress to study in detail the kinetic behavior of the MW-driven polymerization of cyanide.

In addition, the pathways proposed to explain the formation of NH4CN polymers in aqueous media suggest collateral reactions during the course of this precipitation polymerization, such as deaminations, denitrogenations, dehydrocyanations, hydrolysis and oxidations. These last processes are responsible for the incorporation of oxygen in the macrostructures [4], which would lead to a gain of weight considering the total initial mass input in the system (the initial weight of NaCN plus NH4Cl). Figure 2b shows the weight balance found in both series, considering the reaction conditions shown in Table 1, based on the following formula: Weight Balance (%) = [(weight NaCN + weight NH4Cl) − (gel fraction weight + freeze-dry sol content)]/(weight NaCN + weight NH4Cl)]·100. In any case, no direct relationship could be found between the conversion values and the weight balance, as can be observed when comparing Figure 2a,b. In series 1, an increasing weight loss was found as the temperature increased; however, in series 2, when the polymerization time was kept constant, a dissimilar behavior was observed. In this case, at low temperatures, the weight loss, around 5%, was maintained throughout these polymerizations. However, the MW reactions at 170, 190 and 205 ◦C occurred with substantial weight gain, close to 15% or higher. The increase of weight might be related with a minor loss of nitrogen and/or to a higher increase of mass by oxygenation, but the results of the elemental analysis were inconsistent with this proposal (see below), which suggests that complex secondary processes have no reflex directly in the generation of the gel fraction macrostructures.

**Figure 2.** (**a**) Conversion degree, α (%), for the insoluble polymers synthesized according to the reaction conditions shown in Table 1. Above each column is indicated the number of the corresponding polymer. Note that the polymer **2** is included in both series. The conversions for polymers **7** and **9** were calculated using the values obtained in seven independent experiments; for the rest of the polymers, at least three independent syntheses were carried out. (**b**) Weight balance for the global NH4CN polymerization reactions indicated in Table 1, according with the following expression: Weigh balance (%) = [(mg gel fraction + mg freeze-dried solution) − (mg NaCN + mg NH4Cl)]/(mg NaCN + mg NH4Cl)]·100. (**c**) Elemental analysis data for the polymers synthetized according to the conditions indicated in Table 1. At least three representative independent samples were measured for each point in the graph; (**d**) corresponding molar ratios. Note that the %O was calculated by the difference. For the C/O and N/O molar ratios, we considered the %O present in the macrostructure after subtracting this percentage from the water absorbed for the polymers as moisture, which was about ~10% for all of the samples. This amount of water was estimated by the TG data discussed below.

The elemental analysis data displayed in Figure 2c indicates that the C percentage did not present significant differences among the two series under study, and it was about 41%. The same behavior was observed for the %O, at around 17%. Only a slight decrease in the nitrogen content was observed in series 2 with the increase of the temperature, from 41 to 37%. This result was also reflexed in the C/H and C/N molar relationship for this series, when the reaction time was fixed, going from 1.18 to 1.31 for this last ratio (Figure 2d). Therefore, it seems that the increase in the temperature favors the deamination processes during the polymerization based on the lower %N and %H. This fact was in agreement with previous results [13]. In contrast, for series 1, no significant variations were found in the molar relationships with the increasing of the polymerization temperature.

As a result, it can be concluded that a temperature of 170 ◦C seems to be a key parameter in these polymerizations under MW radiation, and also that low polymerization temperatures, such as 130 ◦C, present a singular behavior. However, the compositional data do not give a clear response to the high weight gain observed in series 2.

At this point, it was necessary to point out that the lower yields for the insoluble polymers may be due to the hydrothermal conditions achieved by MW radiation, which favors the oxidation, hydrolysis and decomposition processes of the NaCN and/or other intermediate reaction products, as in Scheme 1, decreasing the availability of them, and therefore reducing the final amount of insoluble polymers collected. Note that all of the reactions were prepared under ambient conditions of pressure and moisture, and neither oxygen or CO2 were removed from the reaction vessel or the solvent.

**Scheme 1.** Possible oxidation and hydrolysis processes likely favored by the MW radiation, which can be undergone by cyanide and by DAMN. These processes lead to lower conversion values for the NH4CN polymerization assisted by MW radiation than for polymerization carried out under conventional heating due to the decrease of the available cyanide and DAMN [20,21].

The left part of Scheme 1 shows the possible processes which can decrease the available amount of cyanide to polymerize. Atmospheric oxygen can lead to the partial oxidation of cyanide at elevated temperatures, generating cyanate. CO2 can react with cyanide to produce carbonate and HCN. In addition, the hydrolysis of the cyanide solutions under heating can produce formiate and ammonium. Under the conditions considered herein, the initial cyanide solutions have a pH of 9.2, and a ~50% of HCN was present in the solutions, which can be hydrolyzed to give formic acid and ammonia. The generation of volatile compounds such as HCOOH, NH3 and even HCN may explain the negative weight balances observed in the NH4CN polymerization above studied. However, the increase observed for the production of insoluble polymers **6**, **7** and **9** might be due to the recycling of the delivered HCN and NH3, generated first as by-products, that finally were implicated in new processes of oligomerization/polymerization, as a closed system was used in the described reactions. On the other hand, the positive weight balances can be due to the formation of no volatile oxygenated molecules such as carbonate and cyanate, as this takes place at high temperatures within series 2. Besides this, in order to explore whether the ambient conditions have a significant influence in these plausible side reactions, where the atmospheric oxygen and the CO2 were implicated, two additional reactions were carried out. Polymers **3** and **7** were again synthetized, but using an inert atmosphere of nitrogen and carefully removing the CO2 and O2 of the water by an N2 stream. The following conversion values were obtained (syntheses were made twice) α(%) = 7.7 ± 0.1 and α(%) = 3.8 ± 0.3 for the analogous polymers **3** and **7**, respectively. Furthermore, in both cases a loss of mass was observed in the total weight balance. Note that the effect of the absence of air in the conversion obtained was outstandingly significant for the analogous polymer **7**. Therefore, the lack of air in the reaction environment does not

seem to prevent the side reactions in the cyanide polymerization assisted by MW radiation; on the contrary, the yield of the reaction was remarkably lower. Due to the great impact of the air in the microwave-cyanide polymerizations, new works will be developed to obtain a full understanding of this factor in these highly complex polymerization reactions.

In addition, we might take into consideration that other hydrolysis and/or oxidations of some the intermediate products during the polymerization process finally lead to the lack of formation of the insoluble polymers, as is the case for diaminomaleonitrile (DAMN). DAMN, the tetramer of the HCN, was considered to be a main intermediate in the generation of the HCN polymers (Scheme 1) (see, e.g., [4]). However, this compound can be oxidized to diiminosuccinonitrile in water, and this is finally decomposed into oxalic acid and urea (upper part of the Scheme 1). Furthermore, DAMN in an aqueous medium can be transformed into formamide, glycine and aminomanolic acid (the bottom part the Scheme 1), as was revealed by Ferris and Edelson [20].

The higher yields (~75%) in the hydrothermal polymerizations of DAMN than were the obtained in analogous NH4CN polymerizations (35–40%) using conventional heating for the production of black insoluble HCN-derived polymers [1], and the significant diminution of these yields in the MW-driven polymerization of DAMN (~30%) (a further work is in progress; no data are shown here) allows us to reaffirm the following points: (i) the DAMN was a main intermediate product cyanide polymerization, as similar—although not identical—final products can be obtained in both ways, i.e., the cyanide and DAMN polymerizations produced resemble—but are not identical—to HCN-derived polymers [1]. The possible polymerization of cyanide by other different DAMN pathways through its dimer and/or its trimer, aminomalononitrile (AMN), for the generation of the suggested C=N networks of the HCN-derived polymers must also be taken in consideration (Scheme 2); (ii) the higher yields obtained for the DAMN polymers seem to indicate that the cyanide undergoes several processes in water, which prevent the generation of DAMN and the subsequent polymerization of this oligomer. Therefore, it seems that the cyanide polymerization via the DAMN pathway was the main process.

As a summary of the above, one can say that: (i) the lack of air in the cyanide polymerization assisted by MW radiation does not seem to inhibit the side reactions which prevent the generation of DAMN, and (ii) the decomposition processes of the DAMN are related to the presence of water. The increase of the temperate increases the ratio of DAMN decomposition in water environments. Both aspects are the main focus of other works in progress.

The FTIR spectra of both series closely resembled those reported from the NH4CN polymerizations under a classical heating and also assisted by MW at 180 ◦C (Figure 3a,b) [1,4,13,14]. In these previous works, analyses of the features for these spectra were discussed in detail. The greatest difference between the spectral features of series 1 and 2 was the relative intensity of the bands related with the nitrile groups (-C≡N), around 2200 cm−1. This difference was clearly observed when the EOR values (the extension of the reaction, EOR = [I1645/(I1645 + I2200)] 100) [22] were represented (Figure 3c). For series 1, this value decreased with the temperature, and by the contrary, for series 2 the behavior was the opposite. This quantitative spectroscopic parameter was directly related to the progress of the polymerization reaction under traditional heating conditions, i.e., the increase of the monomer conversion along the reaction time gives noticeably higher EOR values [22]. However, there was no observed clear relationship between the EOR and the conversion degree, especially in series 2 (Figure 3d, upper part). This behavior was also observed from the NH4CN polymers synthetized at 180 ◦C by MW [13].

**Figure 3.** (**a**) Representative FTIR spectra of polymers **1–5** (series 1), and (**b**) of polymers **2**, **6–9** (series 2); (**c**) relationship of the EOR (extension of the reaction) and the EOC (extension of the conjugation) with the temperature; (**d**) dependence of both the EOR and EOC relationships on the conversion degree. (**c**,**d**) Each point was calculated from at least three FTIR spectra from independent samples.

In addition, the calculation of the EOC (extension of the conjugation, EOC = [I1645/(I1645 + I3330)]·100) [1,13] indicated no significant differences between series 1 and 2, with the exception of the values determined at 170 ◦C (Figure 3c). Likewise, there was no clear trend found when these values were confronted with the conversion. Again, these spectroscopic data revealed that 170 ◦C shows a remarkable singularity with respect to the rest of the temperatures under study.

Simultaneous thermal analysis (STA) for representative samples of polymers **1–9** was carried out (Figure 4), and some parameters obtained from these measurements are listed in Table 1. All of the samples showed the same amount of moisture, at about 10%; however, while the NH4CN polymers from series 1 presented a similar content of char residue, 19–22%, the samples from series 2 exhibited an increase of the char from 13 to 25% with the increase of the temperature. Polymer **9** was the sample that presented the highest percentage of char, and by the contrary, polymer **6** registered the lowest value. In general, a greater amount of char was related with highly cross-linking macrostructures and rigid chains; in particular, NH4CN polymers were associated with the presence of cross-linking oxygenated groups, such as inter- and intramolecular amide bonds [4]. The C/H, C/N, C/O and N/O molar ratios were 0.9, 1.2, 4.1 and 3.5, respectively, for polymer **6**; and 1.0, 1.3, 3.5 and 2.7, respectively, for polymer **9**. Thus, it seems that the higher thermal stability of sample **9** was related with greater oxygenated macrostructures, and with a minor amount of nitrogenized groups, but not with a more conjugated structure, considering the C/H values and the EOC data. Therefore, it might be proposed that this stability may be due to the presence of oxygenated cross-linking groups.

**Scheme 2.** General pathways for the generation of HCN-derived polymers in water. In this scheme is shown the oligomer formation as far as the tetramer DAMN, as well as proposed macromolecular structures from the dimer (red color), the trimer (AMN) (black color) and DAMN (green color) according to recent works [1,3,4,23]. Note that the incorporation of oxygen in some of the structures for the suggested macromolecular systems was due to hydrolytic processes (blue color).

On the other hand, the DTG curves showed similar profiles to other NH4CN polymers synthetized at 80 ◦C [1]. Five characteristic peaks were observed: (i) at ~70 ◦C (water desorption); (ii) at 220–260 ◦C and at ~400 ◦C (thermal break of the weakest bonds); and (iii) at ~670–680 ◦C and 820–840 ◦C (high thermal decomposition, carbonization stage). Only a slight difference was found for polymer **5** with an additional degradation peak at 477 ◦C, and the decrease of the relative peak intensity at ~260 ◦C in series 2 with the increase of the temperature. In addition, the DSC curves were really dissimilar to those registered for NH4CN polymers synthetized at lower temperatures [1,4], proving that the thermal analysis techniques provide excellent fingerprints to distinguish NH4CN polymers with very similar FTIR spectra [4]. For both series of experiments, the first endothermic peak, at 75–85 ◦C, can be related to the loss of the absorbed water in the polymeric matrix, and the rest of peaks can be related to degradative processes.

Some morphological aspects, such as the hydrodynamic diameter, were measured for representative polymers **1–9** (Z-average, Table 1). Considering this parameter, for series 1, the particle size was decreased with the increase of the temperature, reaching nanoparticles with a diameter of ~280 nm and with relatively low PdI = 0.22–0.25. On the contrary, in series 2, although an apparent decrease in the molecular size may be related with the increase of the temperature, the PdI values were higher than those in series 1, indicating a higher degree of heterogeneity. Indeed, the data from sample **9** were not properly registered due to a presumable aggregation of the particles.

**Figure 4.** Thermal analyses for representative samples of polymers **1**–**9**. (**a**,**b**) TG curves; (**c**,**d**) DTG curves; (**e**,**f**) DSC curves.

The XRD pattern profiles for NH4CN polymers **1–9** have the same look as those previously reported for analogous polymers synthetized using a conventional hydrothermal procedure or MW radiation [13,14] (Figure 5a,b). It was expected that an increase in the temperature would lead to a more ordered macrostructure [13,14]. This is the unique observed peak (2θ ~27◦) related to graphitic-like two-dimensional (2-D) structures, such as layered g-C3N4 [16,24], would be higher and narrower at the higher polymerization temperatures. However, this prediction was not valid for the range of temperatures under study when analyzing the data of the crystallinity of the different samples collected in Table 1. There was no clear relationship between the order level of the macrostructures with the increase of the temperature or the reaction time. All the samples exhibited an average crystallinity values of 65%, with the exception of polymers prepared at 130 ◦C, with values higher and lower values than this; showing a particular behavior of the cyanide polymerization at this temperature at least related with the internal order of the macrostructure.

**Figure 5.** XRD patterns of representative samples from both series, (**a**) and (**b**), synthesized by a MW method.

It was clearly shown above that the heating by MW radiation has an unexpected and significant influence in the behavior of the cyanide polymerization, providing results which were not previously observed with conventional heating, i.e., there was not a proportional direct relationship between the reaction time and the conversion degree (%) and EOC [14,22], or temperature and crystallinity [14], for example. However, the final reaction products present resemble the spectroscopic and thermal characteristics of analogous HCN-derived polymers synthetized using conventional heating; for the size of the particles and the higher order of the macrostructures, this was the main difference found between the conventional heating against MW radiation. No nanoparticles were identified in NH4CN polymers synthetized using conventional heating (the Z-average and PdI could not be measured; the data are not shown; however, see [1] for SEM images); at the lower temperature herein considered, the greater size of particles were observed. In addition, at 130 ◦C and 170 ◦C, particular features were observed with respect to the syntheses carried out at the other temperatures. The global and proper interpretation of all of these results is not simple and trivial. Thus, multivariate analysis could be an excellent analytical tool to reach a better and more objective interpretation of the data presented above, and could help us to choose preferential conditions to tune the nanoscale of the NH4CN polymeric particles. In this way, a PCA was carried out to reach a first full overview of the MW radiation effect in the NH4CN polymerization behavior (Figure 6). For this analysis, the total data for the nine polymers (Table 1) were considered. The conversion degree, the balance weight, and all of the molar ratios were calculated; the EOR, the EOC, the crystallinity, the char content, the moisture percentage, the Z-average and the PdI were taken in consideration, explaining 99% of the variance. The first axis shows the highest positive correlation among a wide number of variables, with the exception of EOC and the balance weight. The right-hand portion of the first axis was predominantly occupied by EOR, moisture, monomer conversion, C/O molar ratio, Pdl, Z-average and N/O molar ratio; the left-hand portion was occupied by crystallinity (%) and the C/N molar ratio. The second axis indicates the highest positive correlation among char (%) and the C/H molar ratio. By contrast, the EOC and balance weight exhibited the highest negative correlation with this axis. From these statistical results, one can say that the there is a direct relationship between the conversion degree and the EOR values, as was expected, although a clear relationship was not observed in Figure 3d. In addition, a higher conversion degree was directly related to less-oxidized structures; interestingly, this minor oxidation degree was directly related with a higher size of the particle and a higher polydispersity degree. On the other hand, the percentage of char (%) was strongly correlated with the molar C/H ratio, indicating that the more conjugated structures are the more thermally stable, but

apparently independent of the %O in the structure. A higher internal order based on the percentage of crystallinity is directly related with the C/N ratio and with the char (%), i.e., the more carbonaceous and more conjugated structures are the more crystalline and also the more thermally stable, but the Z-average and crystallinity (%) are independent variables. Therefore, if the MW radiation is considered to obtain nanoparticles of NH4CN polymers, as under conventional heating these are not formed and MW notably reduces the reaction times, it seems that reaction conditions which lead preferentially to the oxidization of structures but are highly conjugated should be considered.

**Figure 6.** PCA biplot based on the physicochemical variables: C/N, C/O and N/O molar ratios; char (%); crystallinity (%); α (%) conversion; EOC and EOR; balance weight; Pdl; moisture (%); and Z-average (nm) for the polymers obtained. The relative proportions of the different variables are indicated by arrows, and polymers are indicated by dots.

With respect to the polymeric samples, PCA analysis indicated groupings of the NH4CN polymers into well-differentiated sets: (i) one corresponding to series 1 (polymers **2**, **3**, **4** and **5**), with the exception of polymer **1**; (ii) another one with polymer **7** and **8**; (iii) and polymers **1**, **6** and **9** isolated from the rest, indicating a particular behavior, as was expected especially for **1** and **6**, as they were the polymers synthetized at 130 ◦C, although polymer **1** was next to the group of series 1, in agreement with a relative resemblance with the samples of this group. In relation to the first group, the clear grouping of polymers **2**, **3**, **4** and **5** indicated that, statistically, the choice of the equivalent reaction times for different temperatures according to the instruction of the manufacturer of the MW reactor, and considering the experimental results indicated above, gives NH4CN polymers with similar characteristics. The second group, formed by polymers **7** and **8,** showed higher char (%), C/H and C/N molar ratios and crystallinity (%), indicating that these polymers are thermally stables carbonaceous structures. For the first, the polymers **6** and **9,** and to a

lesser extent **1**, showed a higher EOR, moisture (%), α (%), C/O molar ratio, Pdl, Z-average and N/O molar, i.e., they presented the polymeric particles with the highest size and lowest oxygen content. Therefore, the lowest and the highest temperature here studied do not seem adequate for the development of nanoparticles.

These well-defined groupings by similar and resemblance features revealed by the PCA analysis, in a certain way, could also be observed for the shape differences found by SEM (Figure 7). Thus, polymers **2**, **3**, **4** and **5** presented rice-shaped nanoparticles and other stacking oval particles together with isolated nanofibers. Polymers **7** and **8** showed clear groupings of nanofibers together with other particles with other shapes (please see Figures S1–S9 from the supplementary information for details), whereas polymer **9** was estranged from this set in the PCA analysis, showing isolated nanofibers and undefinedshape particles. Polymers **1** and **6**, synthetized at the lowest temperature, presented a particular behavior. Polymer **1** displayed similar rice-shaped nanoparticles and staking oval particles to the rest of the polymers from series 1, but also spherical particles similar to those identified in the control polymer [1]. The particles observed in polymer **6** were similar to the particles found for series 1, although the staking oval particles were not identified. These SEM images also seemed to indicate that 130 ◦C and 205 ◦C are not optimal temperatures to produce nanoparticles from cyanide, which is in agreement with the PCA results.

**Figure 7.** SEM images for NH4CN polymers representative samples from series 1 and 2, Table 1.
