OPERACIONES CON EL CENACE

In recent literature the condensation of melamine has probably become the most important synthetic route for accessing melon. For much of the synthetic work con- ducted as part of this thesis a substance which will, as follows, be referred to as “raw melon” was used. This is, however, not meant to say that “raw melon” discussed in this section is chemically identical with “raw” or “crude” melon mentioned elsewhere in the literature.[q] Raw melon was usually prepared by heating melamine in a crucible loosely covered by a lid (cf. Section 12.1.13 for experimental details). The reaction was typically conducted at a temperature of 490 °C with a reaction time of several days. In accordance with other sources,[74-78,108,153] melon is yielded as a yellow solid (cf. Figure 51). The compound is insoluble in any solvent. Solubility observed under very harsh condition like prolonged treatment with hot acids or bases is always due to decomposition (cf. Chapters 7, 9). Raw melon displays high thermal stability and can temporarily withstand temperatures up to 600 °C. No traces of melting or phase transitions could be detected prior to the decomposition by DTA/TG experiments (cf. Figure 52). This thermal stability is, however, only observed for limited durations of heat ex- posure. In open systems, preparative amounts of melon decompose after several weeks at temperatures of 500 °C or higher.

Figure 51. A typical sample of raw melon.

Elemental analysis of samples of raw melon shows good agreement with the com- position of the idealized polymer. It has not been possible to significantly increase the condensation grade by heating at higher temperatures and / or prolonged reaction times (cf. Table 18). An increasing tendency towards total decomposition is observed in- stead. This decomposition, however, takes place rather slowly as was already men-

tioned in the preceding section. 0 200 400 600 800-12

-10 -8 -6 -4 -2 0 0 200 400 600 800 -7 -6 -5 -4 -3 -2 -1 0 1 T G / m g H e a t flo w / µV Temperature / °C Exo

Figure 52. Thermal stability of a typical sample of raw melon measured by DTA/TG. Heat flow is drawn in red TG is drawn in black. Initial weight =

11.57 mg, scanning rate = 5 °C min-1_.

[q] Similar expressions are often found in the literature but refer to compounds prepared differently. J. v.

Liebig has e.g. termed the residue of the calcination of ammonium thiocyanate “rohes Mellon” which would translate into “raw melon. The spelling “mellon” instead of melon was very common at that time.

6.2 Synthesis by Condensation of Melamine

Table 18. Elemental analysis (combustion) for typical samples of raw melon prepared by condensation of melamine.

C / wt.% N / wt.% H / wt. /

melon prepared at 490 °C[a] _{33.22 62.67 1.83}

melon prepared at 550 °C[b] _{33.55 62.44 1.85}

calcd. for (C6N7)NHNH2 35.83 62.67 1.50

[a] Heated for 4 d;

[b] Heated for 6.5 d.

The highly crystalline melon samples used for the structure elucidation of the polymer in the literature were prepared at reaction temperatures of up to 620 °C.[39] Some experiments aiming to accomplish comparable crystallinity at lower reaction temperatures were conducted as part of this thesis. Very long reaction times were chosen based on the observations made for the melamine melem adduct phases (cf. Sec- tion 3.1). Raw melon was mixed with small amounts of melamine and/or NH4Cl as to

provide a certain ammonia pressure upon heating. Thus a certain amount of reversibil- ity of the condensation reaction was assured favoring good crystal growth. The NH3 pres-

sure was, however much lower than for com- parable experiments starting from pure melamine promoting the condensation at lower reac- tion temperatures. Typical reaction conditions involve heating to 490 °C for about one to two days (cf. Section 12.1.14). 4000 3500 3000 2500 2000 1500 1000 70 75 80 85 90 95 100 Tr ansm is s ion / % Wavenumber / cm-1

Figure 53. FTIR spectrum (ATR) of a typical sample of raw melon.

FTIR spectroscopy was regularly conducted for samples of melon. The spectra display the typical absorption bands of melon in most cases (cf. Figure 53). Rather broad absorption bands corresponding to the N-H stretching vibrations are found between 3300 and 3000 cm-1. A large variety of C-N vibrations are found between 1700 and 1100 cm-1 and the characteris- tic vibration for C3N3 / C6N7 rings is observed around 810 cm-1. Raw melon (cf. Figure 53)

shows a spectrum identical to the ones of samples that had been annealed in the presence of NH4Cl / melamine. In consequence, the method unfortunately is not capable of resolving any

sort of fine structural differences as the spectra of all samples of melon prepared are virtually not discernable. The existing structural and chemical differences for different samples of melon prepared from melamine in this work hence are hardly substantial by nature.

Raw melon displays low overall crystallinity. Only a very broad reflection around 2θ = 28° can be observed. For the depicted sample of raw melon a value of 27.4° corresponding to a distance of 3.3 Å was found. This can be used as a first rough estimation of the inter-layer distance. Upon heating with melamine / NH4Cl

the peak position has slightly shifted to 27.8° which is corre- sponding to an inter-layer dis- tance of 3.2 Å. This distance is identical to the one reported for melon prepared at higher tem- peratures.[39,108] More impor- tantly, however, the reflection is also significantly narrowed. The treatment of raw melon in glass ampoules can thus be definitively

associated with the formation of a much more defined product and a notable increase in crys- tallinity. Nevertheless, high-temperature samples prepared from melamine have shown higher crystallinity than comparable samples prepared by annealing raw melon.

10 15 20 25 30 35 40 45 50 750 1000 1250 750 1000 1250 1500 In te nsi ty / C ount s 2 Theta / °

Melon Heated with Melamine/NH4Cl

Raw Melon

Figure 54. PXRD pattern (Cu-K radiation) of a typical sample of raw melon (black, top) in comparison to a fraction of raw melon subsequently heated with small amounts of melamine and ammo- nium chloride (button, red). A drastic decrease in FWHM of the re- flection around 28° is observed.

Figure 55. Typical electron diffraction patterns (200 kV) exhibiting a significant degree of disorder as observed for raw melon.

10 nm 10 nm

Electron crystallography has been proven as an invaluable instrument during structural investigations of melon[39,108] and PHI.[17] Electron diffraction and TEM investigations were employed to analyze samples of both raw and annealed melon. All samples of melon were

6.2 Synthesis by Condensation of Melamine

prepared in accordance to the statements made in the preceding paragraphs and in Sec- tions 12.1.13 – 12.1.14. The measurements were conducted by Dr. M. Döblinger (LMU München). The data are compared in reference to the abovementioned preliminary investiga- tions.[17,39,108]

The ED experiments conducted for raw melon do not show a significant degree of well- crystalline material (cf. Figure 55). This is not surprising since a low overall crystallinity has also been found in powder XRD as is described above. A large amount of disordered regions are found (cf. Figure 55). A pseudo hexagonal pattern was usually observed in these cases. From a structural point of view, raw melon certainly is no pure compound. Amorphous re- gions are also regularly found. Good crystals displaying a non-disordered diffraction pattern are extremely rare but were sometimes observed nevertheless. Small amounts of a new well crystalline side phase were observed (cf. Figure 56). No definite statements concerning the nature of this compound can be made apart from not being identical with crystalline melon as reported in the literature

Figure 56. Typical electron diffraction pattern (200 kV) of an unknown crystalline side phase ob- served in a sample of raw melon.

Figure 57. TEM image of raw melon after preparation for electron diffraction (copper grid).

10 nm 200 nm

TEM images (cf. Figure 57) reveal some hints at crystalline structures as well as some resemblance to the platelets already observed for melon in the literature.[39,108] This finding is, however, not sufficient in order to draw further structural conclusions. The annealed melon sample showed a much more defined morphology and platelet-like structures are typically observed (cf. Figure 58).

Figure 58. TEM images of a sample of melon after preparation for electron diffraction (copper grid). Platelet- like nanostructures can be observed.

200 nm 500 nm

Electron diffraction patterns (cf. Figure 59) usually match the expectations for the re- ported structure of melon. Various degrees of disorder could be observed for the studied sam- ple, revealing amorphous as well as crystalline regions. In agreement to the results already obtained by powder XRD the overall crystallinity of the sample is, however significantly higher than is the one of raw melon. No hints at the new side phase previously observed dur- ing the study of raw melon were found in the annealed sample. Poly heptazine imide (PHI) was also not observed. Thus it can be rationalized that a low temperature synthesis of PHI from mixtures of melamine and raw melon is probably not a very promising approach. Im- proved synthetic procedures for this compound thus still remain to be devised.

Judging from its chemical reactivity and simple spectroscopic properties raw melon cer- tainly is a heptazine-based polymer showing some resemblance with reported crystalline melon. Raw melon must be considered a highly disordered version of the melon reported by

Lotsch et al.[39,108] Especially turbostratic disorder of the H-bonded layers can probably be expected. This is consistent with the findings from powder XRD, indicating a slight increase in inter-layer distance, and the fact that the IR spectrum remains virtually identical for all dif- ferent forms of melon. A significant change in H-bonding interactions would presumably have resulted in a noticeable effect on some absorption bands, the presence of similar H- bonded layers seems plausible. The highly crystalline melon prepared by annealing raw melon with melamine and / or NH4Cl is structurally identical to the melon reported in the literature

as could be shown by ED. The behavior of raw melon upon annealing shows that the linear polymer melon, displaying its reported structure,[39,108] probably is another favorable thermo- dynamic sink. This can in a way be considered similar to the findings concerning melamine or melem, which are also regularly favored under equilibrium conditions. Another interesting finding is, that the annealed melon samples usually only show a slight yellow color, actually much less intense than the one observed for raw melon. Thus the coloring of melon probably is owed to defect sites and beginning carbonization to a much higher degree than it is to the properties of the actual (idealized) polymer itself.

6.2 Synthesis by Condensation of Melamine

200 nm 200 nm

Figure 59. Exemplary ED patterns (200 kV) observed for a sample of melon heated with melamine / NH4Cl.