Synthesis of AgN5 and its extended 3D energetic framework

The pentazolate anion, as a polynitrogen species, holds great promise as a high-energy density material for explosive or propulsion applications. Designing pentazole complexes that contain minimal non-energetic components is desirable in order to increase the material’s energy density. Here, we report a solvent-free pentazolate complex, AgN5, and a 3D energetic-framework, [Ag(NH3)2]+[Ag3(N5)4]ˉ, constructed from silver and cyclo-N5ˉ. The complexes are stable up to 90 °C and only Ag and N2 are observed as the final decomposition products. Efforts to isolate pure AgN5 were unsuccessful due to partial photolytical and/or thermal-decomposition to AgN3. Convincing evidence for the formation of AgN5 as the original reaction product is presented. The isolation of a cyclo-N5ˉ complex, devoid of stabilizing molecules and ions, such as H2O, H3O+, and NH4+, constitutes a major advance in pentazole chemistry.

T he pentazolate anion, cyclo-N 5ˉ, has recently been stabilized as (N 5 ) 6 (H 3 O) 3 (NH 4 ) 4 Cl 1 and Co(N 5 ) 2 (H 2 O) 4 · 4H 2 O 2 . This discovery has received much attention due to the potential applications of cyclo-N 5ˉi n high-energy density materials (HEDMs) and as a starting material for the syntheses of inorganic ferrocene analogs. However, these cyclo-N 5ˉc omplexes contained non-energetic counter ions or groups to enhance their stability, thus impacting their energetic properties. The successful synthesis of an essentially naked cyclo-N 5ˉs alt still has a huge challenge for the fascinating pentazole chemistry and related materials science.
HEDMs require both low sensitivity and high performance 3 . Polynitrogen compounds hold great promise due to their fast energy release and eco-friendly decomposition products [4][5][6][7] . Major advances in this area have been made during the past two decades, the two most remarkable new species discovered in this field are the pentazenium cation, N 5 + , [6][7][8] and the pentazolate anion, cyclo-N 5ˉ9 -12 . However, the reported N 5 + and cyclo-N 5c omplexes generally contain non-energetic counter ions or groups to enhance their stabilities. For example, SbF 6ˉo r SnF 6 2ā re non-energetic counter ions in N 5 + salts 13,14 , and H 2 O, Clˉ, NH 4 + , and H 3 O + are used to stabilize the cyclo-N 5ˉa nion 1 . These non-energetic components impact their energetic properties, such as heat of formation and detonation parameters. Therefore, it is important to reduce or eliminate these nonenergetic components.
As part of our long-continued research, here, we report the synthesis of a water-stabilized cyclo-N 5ˉs alt, [Mg(H 2 O) 6 ] 2+ [(N 5 ) 2 (H 2 O) 4 ] 2ˉ, in which the non-energetic Clˉof (N 5 ) 6 (H 3 O) 3 (NH 4 ) 4 Cl was removed. For the elimination of the water, a silver cyclo-N 5ˉc omplex (AgN 5 ) was precipitated by the addition of AgNO 3 4 ] 2ˉs olution. By treatment with NH 3 ·H 2 O, this AgN 5 complex was converted to a 3D-framework [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉsalt, which was characterized by its crystal structure. The AgN 5 complex is stable up to 90°C, is photolytically unstable decomposing to AgN 3 and N 2 , and Ag and N 2 are its only final decomposition products. The isolation of a silver cyclo-N 5ˉc omplex, devoid of stabilizing molecules and ions, such as H 2 O, H 3 O + , and NH 4 + , constitutes a major advance in pentazole chemistry.

Results
Materials synthesis and structural design. The schematic in Fig. 1 4 ] 2ˉi s an effective step to get rid of the Clˉpresent in the original (N 5 ) 6 (H 3 O) 3 (NH 4 ) 4 Cl salt. The crystal structure of the above Mg salt was determined by single-crystal X-ray diffraction (Fig. 2a, b), which showed that it crystallizes in the triclinic space group P-1. The magnesium center is coordinated to six water molecules in an octahedral fashion with no direct bonding interaction between Mg 2+ and cyclo-N 5ˉ, in contrast to the structure of Co(N 5 ) 2 (H 2 O) 4 ·4H 2 O, where the cobalt ion acts as a shared center linking two pentagonal N 5ˉr ings through two σ-bonds. Interestingly, the cyclo-N 5ˉr ing is surrounded by five crystallographically independent H 2 O molecules, forming the water-stabilized cyclo-N 5ˉs alt. Each of these bridging water molecules acts as an H-bond donor for a nitrogen atom of cyclo-N 5ˉ. Such a coordination mode for cyclo-N 5ˉi s unique and is of vital importance for the further construction of novel pentazolate complexes, because the stability of the water-stabilized cyclo-N 5s alt is determined primarily by hydrogen bonding. These hydrogen bonds can be relatively easily broken, if other cations can trap the cyclo-N 5ˉa nion by forming strong chemical bonds. These considerations have sparked our interest in the preparation of other novel cyclo-N 5ˉc omplexes. As a consequence, we synthesized the AgN 5 complex and its 3D-framework 4 ]ˉwas determined by single-crystal X-ray diffraction analysis. It crystallizes in the monoclinic space group P2 1 /c with a calculated density of 3.2 g/cm 3 at 123 K (Supplementary Tables 2-6). The density value is the highest crystal density reported so far for any cyclo-N 5ˉc omplex 15 , and is largely due to the presence of four heavy silver atoms. As depicted in the Oak Ridge Thermal Ellipsoid plot (ORTEP) of [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉ (Fig. 3a), the asymmetrical unit contains half of an [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]m olecule, which is composed of two Ag(I) cations (50% occupancy for Ag1 and Ag3, 100% occupancy for Ag2), two cyclo-N 5ˉr ings, and one coordinated NH 3 molecule. One cyclo-N 5ˉr ing is no longer perfectly planar, showing a small degree of distortion, as evident from the torsion angles of N(6)-N(7)-N(8)-N(9) being -0.3°and N(8)-N(9)-N(10)-N(6) being 0.2°. In contrast, the other cyclo-N 5ˉr ing resists distortion from planarity, causing a change in the N-N bond lengths (1.323-1.336 Å), which are slightly longer than the N-N bonds (1.318-1.320 Å) in Figure 3b shows the coordination environment of the Ag cations. There are three crystallographically independent Ag centers in the structure. Ag3 is bridging between two ammonia molecules in a linear configuration with relatively short Ag3-N distances (2.110 Å; N(11)-Ag(3)-N(11), 180°). Ag2 is coordinated by four cyclo-N 5ˉr ings, where four N atoms (N1, N4, N6, N9) adopt a distorted tetrahedral configuration around Ag2, with intermediate Ag2-N distances ranging from 2.332 to 2.370 Å, whereas Ag1 is surrounded by three pairs of cyclo-N 5r ings (N3, N7, and N10) adopting an octahedral geometry with two cyclo-N 5ˉr ings at the apical positions and four cyclo-N 5r ings at the equatorial sites. The average Ag1-N distance of 2.519 Å is much longer than the reported values for triazole complexes of Ag(I) (average 2.11 Å) 16 , and the longest bond in the structure, Ag1-N10 (2.669 Å), indicates that the interaction between Ag and cyclo-N 5ˉi s weak.
Physicochemical properties. The [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉ3D framework was further investigated by X-ray photoelectron spectroscopy (XPS). Figure 4a shows the XPS wide scan spectrum, which exhibits N1s and Ag3d peaks only. Two peaks at 368.58 and 374.48 eV generated by photoelectrons emitted from the Ag3d core level, can be observed (Fig. 4b), which indicate the presence of only one type of oxidation state for silver that coordinates to the nitrogen atoms in cyclo-N 5ˉa nd NH 3 . Figure 4c    (Fig. 5a), two new characteristic bands are observed at 393 and 3266 cm −1 , which are due to the symmetric Ag-N 2 stretching mode of [NH 3 -Ag-NH 3 ] + and the NH 3 stretching modes, respectively 23,24 . The infrared spectra of the two compounds show the characteristic absorption of the pentazole rings at ca. 1225 ± 10 cm −1 that is generally present in pentazole complexes. The assignments for the NH 3 25,26 . In the infrared spectrum of the AgN 5 complex one additional unassigned band is observed at 1704 cm −1 . In the vibrational spectra of the AgN 5 complex (Fig. 5b), bands due to N 3ˉa re observed at 2085, 1335, and 604 cm −1 in the RA spectrum ( Supplementary Fig. 3), and at 2016 and 1361 cm −1 in the infrared (IR) spectrum which are due to N 3ˉ( ref. 27 ). The fact that the vibrational spectra of the AgN 5 complex essentially show only bands due to N 5ˉa nd N 3ˉl ends further support to our identification of this compound as a mixture of solvent-free AgN 5 and AgN 3 . This conclusion is further supported the crystal structure of [Ag(NH 3 ) 2 ][Ag 3 (N 5 ) 4 ], in which no evidence for solvate methanol or water molecules was found. Furthermore, the elemental analysis shows the carbon content in the sample of [Ag(NH 3 ) 2 ][Ag 3 (N 5 ) 4 ] to be lower than 0.5%. If some disordered small molecules, such as methanol, existed, they would result in the carbon content to be higher than 0.5%. In addition, no characteristic absorption bands of H 2 O or CH 3 OH were observed in the IR and RA spectra.
The minor slope in the TG curve before 100°C in the Supplementary Fig. 4 can be attributed to the small sample size and some slight decomposition due to the light-sensitivity of the sample. It is also worth mentioning that there were no endothermic peaks in the differential scanning calorimetry (DSC) curve in the 50~90°C temperature region, as would be expected for the evaporation of H 2 O or CH 3 OH. The thermal-decomposition behavior and the stability of Ag (NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉwere investigated by thermogravimetric differential scanning calorimetry (TG-DSC) under an argon atmosphere. [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉshowed a two-step rapid decomposition beginning at 90°C with a mass loss of about 25% between 90 and 134°C, followed by the loss of another approximately 25% between 134 and 320°C ( Supplementary  Figs. 4 and 5). Using thermogravimetric analysis, coupled with mass spectroscopy (TG-Mass), a change of the MS curve at mass 17 (NH 3 ) was observed along with the release of N 2 in the first stage of the decomposition (Supplementary Fig. 6). The second step probably involves the decomposition of AgN 3 to give Ag and N 2 28 . To confirm the overall decomposition process, the decomposition residue from the first weight loss was investigated by slowly heating the complexes under argon to 100°C and then cooling them back to room temperature, followed by IR and powder X-ray diffraction (XRD) analyses. The IR spectrum of the eaks. An additional peak at 3320 cm −1 was assigned to HN 3 29 , suggesting the generation of HN 3 during the first stage of the decomposition, followed by its absorption on the surface of AgN 3 . In the XRD analysis (Fig. 6h), the position and relative intensity of all diffraction peaks match well with those from a standard AgN 3 sample, further confirming the composition of the first-step residue as AgN 3. The XRD powder pattern of the decomposition residue (Fig. 6h) is distinct from that of the original pattern of the starting material before decomposition (Supplementary Fig. 7). One major difference between these complexes and the previously reported (N 5 ) 6 (H 3 O) 3 (NH 4 ) 4 Cl or Co(N 5 ) 2 (H 2 O) 4 ·4H 2 O is that during the decomposition silver particles are produced along with complete release of N 2 . The final thermal-decomposition residue from [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉwas verified by optical microscopy as pure Ag, which has brilliant metallic luster and an irregular, faceted structure ( Supplementary Fig. 8). We have further confirmed this result by using scanning electron microscopy (SEM) and energy dispersive X-ray spectrometry (EDX) to characterize the morphology and determine the chemical phases. Figure 6a indicates that the Ag formed from the thermaldecomposition process consists of multiple nano-layers. Each nano-layer is formed by silver nanoparticles (Fig. 6b), which have small crystallites as evidenced by the XRD analysis. The corresponding intensities of all diffraction peaks are weak due to the relatively low degree of crystallinity (Fig. 6d). The EDX spectrum shows that Ag is the only element detected in the selected region (Fig. 6c), The EDX mappings (Figs. 6e-g) recorded in the whole SEM image indicate that the element on the surface is Ag. By contrast, nitrogen is not observed in the sample region, suggesting the absence of nitrides on the Ag surface. The structure of the AgN 5 complex was also studied in more detail. The XPS wide scan spectrum of the AgN 5 complex showed no significant changes compared to that of [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉ, indicating a similar chemical composition (except for hydrogen). The core-level spectra of N1s, and Ag3d are presented in the Fig. 4a. The only difference between the AgN 5 complex and [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉis that the N1s core levels are centered at 401.08 and 401.28 eV, respectively, which illustrates that the presence of different types of nitrogen groups in the AgN 5 complex has resulted in a slight shift. The IR and RA spectra (Fig. 5b) show only the characteristic peaks of cyclo-N 5ā nd AgN 3 . To explain the formation of AgN 3 , a sample of the AgN 5 complex was exposed to light for 24 h, and then the IR spectrum was re-recorded. It was found that the AgN 5 complex is extremely sensitive to light and completely decomposes to AgN 3 , while the [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉsalt is photolytically less sensitive due to the stabilization effect by the 3D framework. In combination with the structure of [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉand the aforementioned data, it, therefore, can be concluded that the AgN 5 complex is composed of AgN 5 and AgN 3 . This conclusion was further supported by elemental analysis. The total silver content was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES). The found silver content in the AgN 5 complex was 62.3 wt%, intermediate between 60.7% (theoretical silver content in AgN 5 ) and 72% (theoretical silver content in AgN 3 ). The nitrogen content of another sample was also found to be intermediate between the theoretical values for AgN 5 and AgN 3 . Furthermore, the thermal stability and decomposition behavior of the AgN 5 complex were also compared to those of [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉ. As shown in Supplementary Fig. 9, the TG curve also shows two decomposition stages. The first stage involves loss of N 2 from AgN 5 at 120°C to give AgN 3 , and the second stage comprises the complete decomposition of AgN 3 at 337°C to metallic Ag and N 2 .

Discussion
Our results demonstrate the successful syntheses of a solvent-free silver cyclo-pentazolate complex and [Ag(NH 3 4 ] 2ˉ/ AgNO 3 reaction is AgN 5 , which subsequently undergoes partial photolytical and/or thermal-decomposition to AgN 3 . Although we could not obtain a crystal structure for AgN 5 , the indirect evidence for its formation is convincing. The isolation of a cyclo-N 5ˉm etal complex, devoid of stabilizing molecules and ions, such as H 2 O, H 3 O + , and NH 4 + , constitutes a major advance in cyclo-pentazolate chemistry.

Methods
General information. Caution! Solid silver azide and pentazolate are highly energetic and shock and friction sensitive. They should be handled only on a small scale with appropriate safety precautions, i.e., safety glasses, face shields, heavy leather gloves and jackets, and ear plugs.
Materials characterization. All reagents and solvents used were of analytical grade. (N 5 ) 6 (H 3 O) 3 (NH 4 ) 4 Cl was produced according to the methods described in the literature 1 . Fourier-transforminfrared spectra were recorded on a Thermo Nicolet IS10 instrument. Raman spectra were measured with a Renishaw (inVia) Raman spectrometer (785 nm excitation). TG-DSC-mass spectrometry (MS) measurements were performed on a Netzsch STA 409 PC/PG thermal analyzer at a heating rate of 5 K/min under argon atmosphere. X-ray photoelectron spectra (XPS) were carried out on a RBD upgraded PHI-5000C electron spectroscopy for chemical analysis (ESCA) system (Perkin Elmer) with Mg Kα radiation (hν = 1486.6 eV). The crystalline structure was characterized by X-ray powder diffraction (XRD) with a X-ray diffractometer (D8 advance), using a monochromatized Cu target radiation source. The SEM mapping was observed under SEM (FEI verios 460).  4 ] 2ˉ( 0.3 g, 0.87 mmol) in methanol while stirring at 20°C for 30 min, producing the silver pentazolate complex as a pale solid. It was quickly dissolved in 10 equiv. of NH 4 OH and stirred at 0°C for 20 min, followed by warming to room temperature to liberate NH 3 , providing the target product, [Ag(NH 3 ) 2 ] + [Ag 3 (N 5 ) 4 ]ˉ, in 80% yield as an air-stable white solid.
Data availability. The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information