Silver ferrite: a superior oxidizer for thermite-driven biocidal nanoenergetic materials

Silver-containing oxidizers are of interest as biocidal components in energetic application such as thermites due to their biocidal agent delivery. In this study, AgFeO2, was evaluated as an oxidizer in aluminum-based thermite system. This novel oxidizer AgFeO2 particles were prepared via a wet-chemistry method and its structure, morphologies and thermal behavior were investigated using X-ray diffraction, scanning electron microscopy, thermogravimetric analysis and differential scanning calorimetry, and time-resolved temperature-jump time-of-flight mass spectrometry. The results indicate the decomposition pathways of AgFeO2 vary with heating rates from a two-step at low heating rate to a single step at high heating rate. Ignition of Al/AgFeO2 occurs at a temperature just above the oxygen release temperature that is very similar to Al/Fe2O3 and Al/CuO. However, with a pressurization rate three times of Al/CuO, Al/AgFeO2 yields a comparable result as to Al/hollow-CuO or Al/KClO4/CuO, with a simpler preparation method. The post combustion products demonstrated that the Al/AgFeO2 thermite reaction produces a fine dispersion of elemental nanosized silver particles which coats the larger alumina particles and is thus bioavailable.

As to silver, it has been pointed out that silver exhibits biocidal properties in the forms of metallic Ag particles and silver ions in a humid environment. 45 Since most of the products of thermite combustion tend to be primarily in the condensed phase, 2 metallic silver particles are the focus of this work. As indicated by Morones et al. 46 and Smetana et al., 47 small particle sizes are necessary for silver particles to perform well in biocidal activities. To deliver not only a high thermal event but also a large amount of small silver particles as the active biocidal sites is the goal of this work.
When it comes to silver-containing oxidizers, Ag 2 O was the obvious choice to be considered as an oxidizer in an aluminumbased thermite system. In 2010, Clark et al. 48 investigated the combustion performance and biocidal abilities of both Al/I 2 O 5 and Al/Ag 2 O thermites using a homemade biocidal reaction chamber. They concluded Al/I 2 O 5 thermite exhibited signicant spore neutralization owing to the generation of lot of iodine gas. In 2011, Russell et al. 49 further studied the ame propagation behaviors of Al/I 2 O 5 and Al/Ag 2 O thermites using mechanical impact and thermal initiation. The results show that Al/Ag 2 O features a lower average ame propagation by about 2.5 times in thermal ignition tests but produce much more energy than Al/ I 2 O 5 in impact-driven ignition tests. They also argued that the energy release of the thermite reactions is signicantly enhanced by reducing the sizes of the oxidizers particles. In the same year, Sullivan et al. 5 investigated the performance of AgIO 3 as an oxidizer in aluminum-based thermite because it decomposed to O 2 , O and I gases when heated at an ultra-high heating rate. Silver was not observed in the mass spectra probably because the temperature was not high enough to reach the adiabatic ame temperature and thus could not vaporize silver. However, Al/AgIO 3 considerably outperformed Al/CuO in pressurization rate due to a large number of gaseous products released from AgIO 3 ; however, its high combustion performance was mitigated by the fact that the reaction products were found to form AgI instead of elemental silver and iodine, thus obstructing its usage in biocidal applications.
Sullivan et al. 2 subsequently synthesized nano-Ag 2 O particles and investigated its reactivity as an oxidizer in biocidal energetic systems since it produces high yields of antimicrobial silver as one of the combustion products. They found that Ag 2 O alone performs poorly in terms of pressurization rate and burn time, but its performance is signicantly improved when combined with one more reactive oxidizer, such as AgIO 3 or CuO. The morphology of the nal products was also studied and indicated that abundant active sites of silver particles were sacriced since some silver particles were trapped within the interior of other products, which might to some extent affect its biocidal activity. Inspired by the fact that CuO addition improves the performance of Al/Ag 2 O signicantly, we embed the extra oxidizer into Ag 2 O molecularly in this work. Since we have some experience on ferrite-type oxides (AFeO 4 ) and delafossite-type oxides (ABO 2 ) previously, AgFeO 2 as a molecularly mixed oxidizer of Ag 2 O and Fe 2 O 3 with a delafossite structure was a good choice. In addition, Fe 2 O 3 has been proven to be a poor oxidizer in Al-based thermite system, 37 it could be of signicant interest if AgFeO 2 which comprises two poor oxides (Ag 2 O and Fe 2 O 3 ) became a strong oxidizer.
A wet-chemistry method was adopted here for the preparation of AgFeO 2 and its thermal behavior were investigated using a low heating rate thermogravimetric analysis and differential scanning calorimetry in an argon environment. Time-resolved temperature-jump time-of-ight mass spectrometry (T-Jump/ TOFMS) was also employed to evaluate the decomposition behaviors of bare AgFeO 2 or Al/AgFeO 2 thermites under rapid heating rates, enabling us to probe the reaction process on a time scale close to that of a combustion event. The results indicate that the decomposition pathways of AgFeO 2 vary in term of heating rates. For a comparison purpose, Al/CuO, Al/ Ag 2 O and Al/Ag 2 O/Fe 2 O 3 were also included in this work. A highspeed camera coupled with T-Jump/TOFMS simultaneously captured optical emission from the ignition/reaction of the thermites allowing us to obtain the ignition time, and corresponding ignition temperature. In addition, constant volume combustion cell tests were performed on aluminum-based thermites. The post combustion products were characterized by X-ray diffraction, scanning electron microscopy, transmission electron microscopy and energy-disperse X-ray spectroscopy. The results demonstrated that the Al/AgFeO 2 thermite reaction produce an enormous amount of nanosized silver particles and feature the best combustion performance in this work.

Materials
The aluminum nanopowders (Al) (Alex, $80 nm) was purchased from Novacentrix. The active Al was 81% by mass, determined by TGA. All metal oxide nanopowders (<50 nm) purchased from Sigma-Aldrich was directly used as received. All the other chemicals were of analytical grade and used as purchased without further treatment.

Preparation of AgFeO 2
AgFeO 2 powders were prepared via a co-precipitation method. For this, 2.66 mmol of Fe(NO 3 ) 3 $9H 2 O and 2.66 mmol of AgNO 3 were dissolved in 20 mL of water and stirred for 30 minutes. The solution was then heated to 80 C and stirred for one hour. Then, 1.5 M of NaOH solution was added dropwise into the solution until its pH reaches 13, followed by another 6 hours of stirring on a hot plate (80 C). The prepared AgFeO 2 powders could be easily isolated from the solution by vacuum ltration and were puried by successively washing with copious amount of distilled water and absolute ethanol. Finally, the product was dried in an oven at 70 C.

Preparation of Ag 2 O
Ag 2 O powders were prepared by adding 0.025 M NaOH solution dropwise into 80 mL of AgNO 3 solution (0.005 M) with stirring until the solution become a grey-yellow colloidal suspension. The suspension was kept at 60 C for another 2 hours to ensure complete reaction. The prepared Ag 2 O powders were collected by centrifugation and washed with distilled water and then absolute ethanol three times. The solid pure Ag 2 O was obtained aer being dried in an oven at 70 C for 10 hours.

Preparation of thermites
Aluminum nanopowders was stoichiometric mixed with AgFeO 2 , Ag 2 O, CuO and Fe 2 O 3 based on the following equations, respectively, in dry hexane followed by 30 minutes sonication (Table 1). Aer room temperature evaporation of the solvent the solid thermite powders were collected.

T-Jump/TOFMS measurement and high-speed imaging
The decomposition of oxides particles was investigated using a custom T-Jump/TOFMS 5 Typically, a $1 cm long platinum wire (76 mm in width) with a thin coating of oxidizer or thermite sample was rapidly joule-heated to about 1200 C by a 3 ms pulse at a heating rate of $10 5 C s À1 . The current and voltage signals were recorded, and the temporal temperature on the wire was measured according to the Callendar-Van Dusen equation. MS spectra were measured every 0.1 ms. The detailed experimental set-up is given in our previous papers. 5,33 To identify the point of ignition a high-speed camera (Vision Research Phantom v12.0) was employed to record the combustion on the wire during heating. Ignition temperatures of thermite reactions in vacuum were measured from the correlation of optical emission from high speed imaging and temporal temperature of the wire and were further analyzed in combination with the temporal mass spectra. Each experiment was repeated 3 times.

X-ray diffraction (XRD) measurement and Rietveld renement
The as-prepared samples were characterized by powder X-ray diffraction. Diffraction pattern was measured using Cu Ka radiation in Bragg-Brentano geometry on Bruker D8 Advance powder diffractometer equipped with incident beam Soller slits, Ni b-lter and LynxEye position sensitive detector. Data were collected from 10 to 90 2q with a step size of 0.01578 and counting time of 1 s per step (total exposure time of 180 s per step). Thermogravimetric analysis/differential scanning calorimetry (TGA/DSC) measurement Thermogravimetric analysis and differential scanning calorimetry (TGA/DSC) was performed using a TA Instruments SDT Q600. The analysis was performed under a 100 mL min À1 argon ow with $1.0 mg samples placed into an alumina pan and heated from room temperature up to 1000 C at a rate of 10 C min À1 in argon atmosphere.

Morphologies and structures characterizations transmission electronic microscopy (TEM)
Transmission electronic microscopy (TEM, JEOL JEM 2100 FEG) and scanning electronic microscopy (SEM, Hitachi Su-70) were used to investigate morphologies and structures of thermites. Elemental distribution in the thermites was analyzed by Energydisperse X-rat spectroscopy (EDS) on both SEM and TEM.

Combustion test
Combustion properties of themites were evaluated in a constant-volume combustion cell, with simultaneous pressure and optical emission measurements. In this study, 25 mg of thermite powders was loaded inside the cell (constant volume, $13 cm 3 ) and ignited by a resistively heated nichrome wire. The temporal pressure and optical emission from the thermite reaction were measured using a piezoelectric pressure sensor and a photodetector, respectively. More detailed information on the combustion cell test can be found in our previous publications. 21,25,30 Each experiment was repeated at least 3 times.

Results and discussions
Synthesis and characterization of AgFeO 2 XRD of as-prepared materials shown in Fig. 1A indicates that silver ferrite was successfully prepared. The particles have an oval shape with particle size of $40 nm based on the SEM image shown in Fig. 1B, and suggests it might be a very good oxidizer in a thermite system due to its small size. 33 The thermal stability of AgFeO 2 was studied with a TGA in Ar at a heating rate of 10 C min À1 . The result shown in Fig. 2 indicates a $4% weigh loss at around 650 C, corresponding to O 2 gas release based on eqn (1). To determine the composition of the remaining material, XRD shown in Fig. 2B indicates the formation of both Fe 2 O 3 and Ag. a The thermodynamic data of AgFeO 2 is unknown. Data is taken from Fischer and Grubelich 50 without taking account of the oxide shell on aluminum. High heating rate decomposition of AgFeO 2 was investigated using TOFMS/T-Jump (3 ms, heating rate $ 5 Â 10 5 C s À1 ). Fig. 3 shows the temporal evolution of O 2 from AgFeO 2 during rapid heating. Oxygen was rst detected at around 685 C which is slightly higher than the onset decomposition temperature under low heating rate TGA. Only one O 2 signal stage is observed indicating the decomposition of AgFeO 2 is a one-step event at high heating rates conditions, as compared to its multistage decomposition behavior at low heating rates as discussed previously.
Time-resolved T-Jump mass spectra from rapid heating of AgFeO 2 at 1.6-2.2 ms is shown in Fig. S1 † to further explore its decomposition process. Apparently, the onset decomposition of AgFeO 2 started at around 1.8 ms (685 C) with the appearance of a small O 2 peak. For mass spectra taken at prior times, H 2 O + , OH + , CO + /N 2 + peaks are attributed to the background. 18 Above the decomposition temperature, no new peak except for O 2 peak is found again suggests a one-step decomposition of AgFeO 2 at high heating rate. CO 2 + was also observed which we attribute to decomposition products of the precursor salt residue on the AgFeO 2 surface. 17

Ignition of Al/AgFeO 2 nanothermite
To evaluate the performance of AgFeO 2 as an oxidizer in Albased thermite, physical mixtures of nanosized aluminum and AgFeO 2 were made, following 30 min sonication of the mixture in dry hexane. Sequential snapshot of Al/AgFeO 2 ignition during high-rate heating under vacuum were captured using a high-speed camera and is shown in Fig. 4. Optical emission from Al/AgFeO 2 reaction was rst observed at 1.662 ms with a corresponding wire temperature of $740 C as a sign of ignition. Multiple ignition points are observed in the prior time and then merge into a large bright ame. One should also notice that, upon ignition, thousands of bright dots were rapidly ejected out from the reactants coated on the Pt wire (even more bright dots appeared when ignited in an argon environment as shown in Fig. S2 †). The burn time of the thermite in T-Jump chamber could be roughly obtained based on the visual ame and is about 0.3 ms, which is almost the same as the value obtained from the combustion cell test discussed below. Ag 2 O was included as an oxidizer in Al-based thermites as a control. As the most common used oxidizer, CuO was also included as a standard reference. Since AgFeO 2 (Ag 2 Fe 2 O 4 ) can be molecularly written as Ag 2 O and Fe 2 O 3 , a binary mixture Ag 2 O-Fe 2 O 3 with 1/1 molar ratio is also included for comparison. Fig. 5 shows the relationship between O 2 release temperature in neat oxides and the ignition temperature of corresponding thermites under vacuum. There is a good correlation between the oxygen release from the bare oxidizer and ignition of Al/ CuO, Al/Fe 2 O 3 and Al/AgFeO 2 . Al/Fe 2 O 3 has a very high ignition temperature due to the poor performance of Fe 2 O 3 as an oxidizer. 24 On the other hand, Al/Ag 2 O thermite ignited at around $660 C, essentially the melting temperature of aluminum. Considering Ag 2 O releases oxygen at around 520 C, it is reasonable to conclude that the ignition of Al/Ag 2 O thermite is limited to the melting phase of aluminum like most Al/metal oxides systems. 19 In fact, the ignition temperatures of the other four samples are all higher than the melting point of aluminum indicates that gaseous oxygen released from the decomposition of oxidizers, was insufficient to ignite aluminum when it is still in the solid phase.
Al/Ag 2 O/Fe 2 O 3 ternary thermite ignited at around 800 C much higher than its oxygen release temperature (520 C). Considering Al/Ag 2 O and Al/Fe 2 O 3 ignited at around 660 and 940 C, respectively, an ignition temperature of about 800 C is reasonable for the ternary system Al/Ag 2 O/Fe 2 O 3 . The Fe 2 O 3 addition is more likely hindering/weakening the ignition/  reaction of the ternary system due to its weak reactivity in Albased thermites. 37 As to Al/AgFeO 2 , it ignited at a temperature close to the oxygen release temperature of the corresponding oxidizer and higher than the melting temperature of aluminum following an ignition mechanism similar to Al/Fe 2 O 3 and Al/CuO. 18,19 Moreover, the fact that the ignition temperature of Al/Fe 2 O 3 is much higher than that of Al/AgFeO 2 implies the oxygen involved with Al/AgFeO 2 ignition come from AgFeO 2 rather than the rst-stage decomposed product Fe 2 O 3 as indicated in TGA. Since only one oxygen release stage appeared at the mass spectra and the violent reaction of Al/AgFeO 2 at 1.7 ms ($780 C) implies decomposition of AgFeO 2 is a one-step process during Al/ AgFeO 2 thermite reaction at high heating rate.

Combustion performance of Al/AgFeO 2 nanothermite
The combustion performance of Al/AgFeO 2 at room atmosphere was studied using a constant volume combustion cell, and the results are summarized in Table 2. The pressurization rate and peak pressure in Al/AgFeO 2 reaction are 6200 kPa ms À1 and 325 kPa, respectively, which are considerably higher than those of nano-thermite reaction of Al/CuO. A direct comparison between the temporal pressure traces of Al/AgFeO 2 and Al/CuO was shown in Fig. 6A in which the rst peak was used to determine the corresponding pressurization rate. Clearly, Al/AgFeO 2 reaction reaches its rst peak with a faster rate and higher peak value compared with Al/CuO. We have claimed in our previous works that gaseous species release from the decomposition of the oxidizer is the main cause for the pressurization, which can occur much earlier than ignition/combustion. 37,51 The fact that AgFeO 2 ($16%) has a lower oxygen weight ratio than CuO ($20%) suggests the O 2 release rate of AgFeO 2 must be higher than that of CuO in order to output such high pressure. In addition, the optical signals peak much later than the corresponding pressure peaks for both Al/CuO ($0.5 ms) and Al/ AgFeO 2 (0.9 ms) implying a similar mechanism. 51,52 Therefore, similar to Al/CuO, 37 the ignition/reaction mechanism of Al/ AgFeO 2 is summarized as follows: AgFeO 2 releases O 2 gas ($690  C) and the aluminum core becomes mobile (633 C); and the mixture ignites ($740 C, Fig. 5) and the generated heat further promote the decomposition of AgFeO 2 to pressurize the system. Moreover, with a pressurization rate three times of Al/CuO, Al/AgFeO 2 yields a comparable result as prior work on Al/ hollow-CuO 25 or Al/KClO4/CuO, 37 which make AgFeO 2 a powerful replacement owing to its simpler and cheaper preparation method.
Sullivan et al. 2 has previously reported that Al/Ag 2 O suffers a poor combustion performance in terms of both pressurization rate and peak pressure and incorporation of a small fraction of CuO into the Al/Ag 2 O system improved its combustion performances extensively to even close to the reactivity of Al/CuO. An explanation was provided by Sullivan et al. that CuO addition can increase the reaction temperature and thus further enhance the performance of Ag 2 O as an oxidizer. Consistent with this reported result, Al/Ag 2 O performs poorly in the combustion cell test due to an early release of O 2 ; 2 however, Fe 2 O 3 addition does not show much improvement to Al/Ag 2 O system as CuO does. Instead of improving, Fe 2 O 3 even weaken the combustion  Fig. 6 Temporal pressure and optical behavior of Al/AgFeO 2 and Al/nCuO.    Table 3 where we can see that both CuO and Ag 2 O should perform signicantly better than Fe 2 O 3 in terms of gas production. And this is consistent with our experimental data shown in The burn time of the thermite reactions measured in the combustion cell are shown in Table 2. Clearly, Al/AgFeO 2 features the shortest burn time which makes it the most violent thermite among the four examined. A direct comparison of the optical emission trace of Al/AgFeO 2 and Al/CuO was shown in Fig. 6B. The Al/CuO reaction has a four-times higher peak optical emission but 0.1 ms longer burn time compare with those of Al/AgFeO 2 . This result indicates that Al/AgFeO 2 is a weaker heat generator than Al/CuO; but reacts more rapidly. In general, AgFeO 2 is the best oxidizer among those four in the aluminum-based thermite system from both pressurization and optical emission perspectives.

Post-combustion-product characterization
It has been pointed out previously that the nature and dispersion of the products from a thermite combustion plays an important role in biocidal applications. 27 XRD evaluation of crystalline product species is shown in Fig. 7 and show no evidence of the parent starting materials, but ve new strong peaks indexed to elemental silver are seen. The fact that no peak corresponding to Al 2 O 3 or Fe was observed might suggest they are both amorphous.
For SEM evaluation of the reaction product a 7 mm Â 5 mm rectangular double-sided carbon tape was placed inside the combustion cell chamber. Two representative SEM images of the product were shown in Fig. 8A and B, there are mainly two populations of product particle sizes. One has relatively larger dimension and another with a dimension as small as $80 nm. 2D elemental mapping using EDS shown in Fig. 8, indicate the larger spherical particles are Al 2 O 3 while the smaller particles are Ag. This was also conrmed with 1D elemental line-scan coupled with elemental analysis shown in Fig. S3. † Most importantly for the application as a biocide, is that the smaller Ag particles randomly decorate the larger Al 2 O 3 particles.
Overall, the intensity of iron signal is the weakest among those four elements under EDS mapping implying the random distribution of iron.
TEM of the reaction product along with an EDS elemental mapping result is shown in Fig. 9. As can be seen, there are two  different types of particles similar to the SEM result. One has a spherical shape with a relatively larger particle size and low contrast; another has a spherical/oval shape with smaller size and dark contrast. In the EDS elemental mapping, red represents silver, blue iron, yellow aluminum and green oxygen. Clearly, the three larger spherical particles are Al 2 O 3 . Since the mixing of red and blue gives purple, it is reasonable to conclude that the silver and iron positions are mostly overlapped. It could mean a formation of Ag-Fe alloy; however, the SEM results indicate the iron signal is much weaker and always appears with the presence of both aluminum and oxygen signals. To clarify this speculation, an EDS line-scan analysis was employed, and the result will be discussed below. It should also be noted that some white pixels were found in the purple area, implying the presence of oxygen (green plus red and blue equals to white). Thus, those silver-and iron-containing particles might be partially oxidized that could be a result from handling the product in air.
As we can see from the EDS line-scan result shown in Fig. 10, excerpt for the small oxygen signal peak at $0.4 mm, the aluminum and oxygen signals are almost synchronous in positions indicating Al 2 O 3 . Like the SEM EDS mapping result, iron is the weakest among all four elements. It overlaps with aluminum and silver, respectively, implying the iron is randomly distributed and excluding the hypothesis of Ag-Fe alloy formation.
Comparing these two representative TEM images Fig. 9A and 10A, the morphologies of the silver particles are quite different. The ones in Fig. 9A are nanosized with spherical/oval shapes and are decorated on/near the Al 2 O 3 surface that suggests silver product had vaporized and re-condensed. 2 This mechanism is benecial to biocidal applications due to the wide distribution of small sized silver particles during the violent combustion event. As to the one in Fig. 10A, particles at the bottom right with oddly shapes clearly underwent some sintering. Those results are consistent to a previously reported sintering reaction mechanism of Al-based thermites. 53,54 However, it seems much less reactants were following the sintering mechanism according to the SEM image shown in Fig. 8 where the majority of silver product has small particle size and distributed randomly.

Conclusions
In this study, the AgFeO 2 was prepared via a wet-chemistry method to yield phase pure $40 nm particles. The decomposition pathways of AgFeO 2 were found to depend on heating rates: decomposition to Ag, O 2 and Fe 2 O 3 at $600 C at low heating rate and direct decomposition $685 C at high heating rates.
The ignition of Al/AgFeO 2 was found to slightly higher than the oxygen release temperature and thus with a similar mechanism to Al/Fe 2 O 3 although ignites at a much lower temperature. Fast video imaging indicates a ne smoke dispersing particle.
Combustion cell test showed that Al/AgFeO 2 outperformed other thermites in maximum pressure, pressurization rate and burn time. Moreover, with a pressurization rate three times and 5% less oxygen content of Al/CuO, Al/AgFeO 2 yields a comparable result as to Al/hollow-CuO or Al/KClO 4 /CuO. The fact that Al/AgFeO 2 and Al/Ag 2 O/Fe 2 O 3 share the exact same elemental compositions but feature the highest and lowest pressurization rate, respectively, indicates molecularly incorporation of Ag 2 O into Fe 2 O 3 outperforms the mechanically mixed Ag 2 O/Fe 2 O 3 when they were employed as oxidizers in aluminum-based thermites. Post combustion products indicate the formation of elemental silver nanoparticles ($<80 nm) decorating larger Al 2 O 3 and is this bioavailable.

Conflicts of interest
There are no conicts to declare.