Stabilization of a reduced crystal phase via
lack of O atoms: certain O-deficient metal oxides can stabilize reduced phases. This occurs when a significant fraction part of the metal atoms is reduced. For example, magnetite Fe
3O
4, which contains one Fe
2+ and two Fe
3+, can be formed from Fe
2O
3 (two Fe
3+) when 1/3 of the Fe
3+-atoms is reduced to Fe
2+. Further reduction in all Fe atoms to Fe
2+ forms the FeO phase, while further reduction to Fe
0-atoms forms the metallic, zero-valent-iron material. Similarly, Cu
2O (SnO) is formed when all Cu
2+ (Sn
4+) atoms in CuO (SnO
2) are reduced to the Cu
1+ (Sn
2+) state.
-
The concept of using an oxygen-lean FSP was pioneered by Grass et al. to produce oxygen-deficient metal-oxide particles
[44][26] by placing the FSP nozzle inside a glove box filled with inert nitrogen and regulating the intake of oxidizing gas as illustrated in
Figure 4b,c. The dispersion gas mixture in the flame can shift from a CO
2/H
2O composition (representing traditional, oxidizing flames, see
Figure 4a) to a CO/H
2/H
2O mixture (under reducing conditions)
[44][26]. Noble metal nanoparticles, including Pt, Au, Ag, and their alloys, can typically be produced even in oxygen-rich FSP, i.e., due to the thermodynamic preference of the metal state vs. the oxide state by the noble metal atoms. However, creating non-noble metals necessitates a reductive environment. When cobalt or bismuth organic precursors
[45][27], such as cobalt(II)- and bismuth(III)-2-ethylhexanoate, are burned in a controlled atmosphere (with O
2 levels less than 100 ppm) and with a high fuel-to-oxygen ratio (see
Figure 4b), it enables the swift production of pure Co and Bi metal nanoparticles, enhancing the conventional flame process. With this experimental setup, Stark et al. have explored the creation of metallic bismuth nanoparticles ensuring no soot formation
[46][28]. While the reducing environment might be beneficial for producing metallic particles on a large scale
[45[27][29],
47], it comes with the risks of incomplete combustion
[1]. In the case where the oxygen supply is further constrained, a fine carbonaceous layer tends to form on these metal nanoparticles
[47,48][29][30]. Using this experimental setup, NiMo nanoalloys
[49][31] and ZnS nanocompounds
[50][32] have been reported.
Figure 4. (
a) Conventional FSP (left) and
our
reducing FSP (right), where the anoxic flame is produced by in situ introduction of reducing dispersion gas, e.g., CH
4. (
b) An anoxic FSP reactor, used by Stark, with the whole reactor enclosed in a glove box filled with an inert atmosphere. By adjusting the gas flow rates, it is possible to achieve highly reduced conditions (O
2 < 100 ppm). Used with permission of Royal Society of Chemistry from
[45][27]; permission conveyed through Copyright Clearance Center, Inc. (
c) Schematic depiction of the step-by-step transformation from precursor to oxide, metal, and carbon-coated metal nanoparticles during the reducing flame synthesis process: Initially, the precursor undergoes evaporation and combustion, resulting in oxide nanoparticles. These particles can then be further reduced to their metallic form by H
2 and CO. Throughout this procedure, the nanoparticles increase in size due to aggregation and sintering. By introducing acetylene, these metal nanoparticles can acquire a carbon coating layer.
Strobel and Pratsinis used an oxygen-deficiency FSP process
[51][33] in order to synthesize Fe
2O
3, Fe
3O
4, and FeO nanoparticles. Their setup featured an FSP nozzle with a metal tube (4 cm in diameter and 40 cm in length) positioned directly above it (as shown in
Figure 4b). Situated 20 cm above the FSP nozzle and angled at 45°, an internal mix spray nozzle was directed downward. This nozzle delivered deionized water at a rate of 10 mL/min, dispersed using 5 L/min of N
2. A different oxygen-deficiency FSP setup for the production of Fe
3O
4 nanoparticles may be the utilization of a laminar, inverse diffusion flame
[52][34]. This method takes advantage of the properties of the inverse flame, created when an oxidizer is injected into a flow of surrounding fuel
[53][35]. Contrary to conventional flame approaches, this setup ensures that the iron particle formation occurs in a predominantly reducing atmosphere. As illustrated in
Figure 5a, the burner features two concentric brass tubes with specific outer diameters, enclosed within an 11.4 cm diameter acrylic chamber. This chamber is crucial for protecting the flame from ambient air, preventing additional particle oxidation and potential secondary diffusion flame formation due to excess fuel reacting with room air. The oxidizer, either pure O
2 or an O
2-Ar mixture, is released from the innermost tube and is encircled by a blend of fuel (methane or ethylene), argon, and iron precursor vapor. A N
2 flow enveloped the resulting inverse flame.
Figure 5. (a) Experimental setup of laminar, inverse diffusion flame stabilized on a burner for the synthesis of magnetic iron oxide nanoparticles with reduced oxidation state. (b) The concept of the novel anoxic FSP, as developed by our lab, for ZrO2−x production. (c) (i) Schematic representation of anoxic FSP reactor used for the synthesis of C@Cu2O/Cu0 nanoparticles. (ii) Anoxic FSP reactor configuration utilized for creating CuO and Cu2O nanomaterials.
Recently,
wresearche
rs have exemplified a novel anoxic FSP process, to engineer ZrO
2–x (see
Figure 5b)
[54][36] and C@Cu
2O/Cu
0 (see
Figure 5c)
[55][37] nanoparticles.
OurThe anoxic FSP concept relies on the combustion of CH
4 in the dispersion gas. This introduces reducing agents that can modify the primary Zr particle by creating oxygen vacancies (V
O). XPS and EPR confirm that the increased dispersion of the CH
4 promotes the formation of oxygen vacancies
[54][36]. A more complicated oxygen-deficiency FSP setup, which includes a dispersion feed consisting of {oxygen (O
2)–methane (CH
4)} mixture, in tandem with enclosed FSP flame with radial N
2, is necessary for the synthesis of non-graphitized carbon/Cu
2O/Cu
0 heterojunction (see
Figure 5c)
[55][37]. The modification in the dispersion gas mixture leads to increased temperatures and generates reducing agents for the controlled phase transformation from CuO to Cu
2O and Cu
0 (see
Figure 5c).
2.2. Double-Nozzle FSP Configuration
In the case of mixed structures, e.g., heterojunctions, core-shell compositions, etc., the application of two FSP nozzles that operate in tandem offers advantages. Typical examples include the cases where a nanomaterial (NP1) and a cocatalytic nanomaterial (NP2) are combined. In the conventional single-nozzle FSP, a single precursor contains both the elements of nanomaterial (NP1) and nanomaterial (NP2) and produces the combined material in a single flame (see
Figure 6a).
Figure 6.
Symmetric and asymmetric DN-FSP configuration for two particle formation regarding the (
i
) atomic, (
ii
) particle, or (
iii
) agglomeration scale.
Double-nozzle FSP entails two independent spray flames, with the precursor of NP1 inserted in a different flame than NP2 (see
Figure 6). This method unlocks several options for independent size control, mixing, and specific deposition for the two nanomaterials by altering the primary geometrical parameters of distance and intersection of the flames. As shown in
Figure 6: (i) At a small flame-intersection distance, where the centers of the flames are in contact, the atoms are in the preliminary stages of crystallization, producing well-mixed particles, tending to be similar to the single-nozzle FSP. In this case, the second flame substantially increases the synthesis overall temperature. (ii) When the intersection occurs after the endpoints of the flames, the materials are well crystallized, resulting in well-mixed primary particles of NP1 and NP2. (iii) At increased intersection distance, the two materials mix at their sintering stage or bigger distances at the agglomeration stage.
Thus, by changing the geometrical disposition of the two flames via the parameters a, b, d, Φ
1, Φ
2, and Z (see
Figure 7b), the symmetrical/asymmetrical DN-FSP configuration offers a versatile technology that allows for the control of composite configurations at different synthesis stages, i.e., at the atomic scale, at the particle scale, or the aggregate’s scale (see
Figure 6).
Figure 7. (a) Schematic example of SN-FSP where two precursors are mixed before being fed to the flame. (b) Geometry parameters of DN-FSP. (c) Example of a symmetrical DN-FSP, used for engineering of La-doped SrTiO3, with surface deposition of CuO. (d) Example of asymmetrical DN-FSP.
Al2O3: DN-FSP was first implemented by Strobel et al.
[59][38], producing in one nozzle Al
2O
3 and in the second nozzle Pt/BaCO
3, thus forming individual Al
2O
3 and monoclinic BaCO
3 nanoparticles. Increasing the internozzle distance delayed flame product mixing, increasing the crystallinity of BaCO
3. In contrast, the single-nozzle process yielded Al
2O
3 particles with amorphous Ba species. The two-nozzle process enhanced NO
x storage behavior, while the single-nozzle approach showed negligible NO
x retention
[59][38]. Following this successful novelty method, a series of Al
2O
3-based articles were published, herein chronologically presented: Minnermann et al.
[60][39] produced in one nozzle Al
2O
3 and in the other pure oxide or mixed CoO
x. Single flame synthesis is inadequate for producing an effective Al
2O
3/Co FT catalyst due to inadequate reducible cobalt oxide support particle size. The DN-FSP geometry significantly influences the resulting catalyst, yielding smaller alumina particles as the intersection distance increases, resulting in good adhesion of the two oxides and good stabilization. Høj et al.
[61][40] produced Al
2O
3/CoMo by DN-FSP, and varying flame mixing distances (81–175 mm) minimized the formation of CoAl
2O
4, detectable only at short flame distances. Notably, employing DN-FSP synthesis achieved superior promotion of the active molybdenum sulfide phase, potentially attributed to reduced CoAl
2O
4 formation, consequently enhancing Co availability for promotion. Schubert et al.
[62][41], through DN-FSP, produced Al
2O
3/Co enhanced with Pt (0.03, 0.43 wt%) deposition in the first nozzle and other materials in the second nozzle. Noble metals enhance catalyst reducibility, yielding abundant metallic Co sites. Due to their high cost, optimizing synthetic strategies for low concentrations is essential. Regardless of the preparation approach, adding 0.03 wt% Pt significantly improves catalytic activity in CO
2 methanation, and 0.43 wt% Pt marginally increases the catalyst reduction. Using DN-FSP, Horlyck et al.
[63][42] produced Al
2O
3/Co with Lanthanum doping (0–15 wt%). Increased La content and wider nozzle distance suppressed undesirable CoAl
2O
4 spinel phase, promoting easily reducible Co species. La addition enhanced carbon resistance, ensuring maximum methane conversions at 15 wt% La without catalyst deactivation or carbon formation. Stahl et al.
[64][43] used DN-FSP to produce Co/Al
2O
3; in the nozzle of Al
2O
3, one additional particle—SmO
x, ZrO
x, or Pt—was formed contributing different cocatalytic effects, enhancing surface hydrogen or carbon oxide concentrations (see
Figure 8a,b). All catalysts had consistent morphology with interconnected 12 nm alumina oxides and
~
8 nm cobalt oxides. For CO
2 methanation, Pt and zirconia proved optimal, aligning with Pt-enhanced H
2 adsorption and zirconia’s higher CO
2 adsorption due to oxide sites with medium basicity.
Figure 8. (a) TEM images revealing the local distribution of cobalt and oxygen for Pt-Al2O3/Co3O4, (b) EDX measurements for chemical composition. DΝ-FSP-prepared (c) SiO2/Co, (d) SiO2-TiO2/Co, (e) and TiO2/Co; left images show STEM-HAADF and right images show EDX mappings of the elements Co (blue), Si (red) and Ti (yellow). (f) Particle size distributions of Co3O4 for the materials SiO2, SiO2-TiO2, and TiO2. (g) STEM-HAADF of the nano-mixed CeO2:Eu3+/Y2O3:Tb3+ and its elemental mapping for Ce in red and Y in green, (h) dTEM distribution of CeO2:Eu3+ and Y2O3:Tb3+.
TiO2: Grossmann et al., through the utilization of DN-FSP, produced TiO
2 with deposited Pt particles
[67][44]. Geometric configurations in DN-FSP strongly influenced Pt particle size and distribution on TiO
2. Larger intersection distances and smaller angles result in nonuniform large and broadly distributed Pt clusters on TiO
2. Conversely, smaller distances and larger angles enhance Pt dispersion and a uniform mixing, akin to single flame; however, DN-FSP allows for individual tuning of compound particle sizes. Solakidou et al. produced {TiO
2-Noble metal} nanohybrids, with deposition of Pt
0, Pd
0, Au
0, or Ag
0 [68][45]. As shown, DN-FSP is superior vs. single-nozzle-FSP for finely dispersing noble metals on TiO
2 support, achieving a narrower size distribution
[50][32]. DN-FSP promoted intraband states in TiO
2/noble metal, reducing the band gap. Efficient H
2 generation presented the following trend: Pt
0 > Pd
0 > Au
0 > Ag
0, in line with a higher Schottky barrier upon TiO
2 contact
[50][32]. Gäßler et al. produced SiO
2, TiO
2, and SiO
2-TiO
2 mixture with DN-FSP deposition of Co
3O
4 (see
Figure 8c–f)
[65][46]: titania, comprising anatase and rutile phases, the SiO
2-TiO
2 mixed support, with separate anatase and silica phases. H
2O adsorption varies significantly based on the support: SiO
2 < SiO
2-TiO
2 < TiO
2. CH
4 formation rate increased with higher TiO
2 fractions, while CO formation rate peaked in the mixed support. Psathas et al. used DN-FSP to engineer heterojunctions of perovskite SrTiO
3 with deposited CuO nanoparticles (0.5 to 2 wt%)
[58][47]. Higher CuO deposition led to larger SrTiO
3 particle sizes due to increased enthalpy from the second flame
[40][22]. Scanning TEM depicted small CuO particles (<2 nm), mainly found on the surface of SrTiO
3. The dopant concentration significantly controlled the selective production of H
2 or CH
4 from H
2O/CH
3OH. CuO incorporation drastically shifted production to CH
4, achieving a rate of 1.5 mmol g
−1 h
−1 for the La:SrTiO
3/CuO catalyst (0.5 wt%)
[58][47].
Other particles: Tada et al., using DN-FSP, produced a ZrO
2/CuO heterostructure
[69][48]. Changing the geometrical parameters of DN-FSP altered the proportion of interfacial sites vs. copper surface sites. As active sites are primarily at the metal–oxide interface, ZrO
2/CuO with smaller CuO clusters exhibited higher activity in methanol synthesis via CO
2 hydrogenation. Gockeln et al., by a combination of DN-FSP and a lamination technique
[73][49], synthesized in situ carbon-coated nano-Li
4Ti
5O
12 Li-ion battery electrodes. Li et al. synthesized LiMn
2O
4 spinel as a cathode material for Li-ion batteries via screening 16 different precursor–solvent combinations
[70][50]. To overcome the drawback of capacity fading, the deposition of AlPO
4 (1–5%) via DN-FSP was homogeneously mixed with LiMn
2O
4. The optimal 1% AlPO
4 with LiMn
2O
4 demonstrated an energy density of 116.1 mA h g
−1 at 1 C (one-hour discharge). Henning et al. used DN-FSP to engineer luminescent biosensors CeO
2:Eu
3+/Y
2O
3:Tb
3+ [66][51]. CeO
2:Eu
3+ nanoparticles (6 nm, 22 wt%) and Y
2O
3:Tb
3+ nanoparticles (32.5 nm, 78 wt%) were shown to function as robust optical-based ratiometric H
2O
2 biosensors (see
Figure 8g,h). Based on the collective effect, H
2O
2 caused significant luminescence quenching in CeO
2:Eu
3+ nanocrystals, but Y
2O
3:Tb
3+ nanoparticles were unaffected
[48][30].
Asymmetric Double Flame: Lovell et al. utilized asymmetric-DN-FSP geometry to control the SiO
2 interaction with Ce
0.7Zr
0.3O
2 nanoparticles
[71][52]. Tuning the intersection distance during DN-FSP (18.5 to 28.5 cm) prevented silica coating. Short intersection distances led to high surface-area silica encapsulating ceria-zirconia, while longer distances suppressed this encapsulation. The material at longer intersection distances, used as Ni support for dry methane reforming, showed enhanced oxygen storage capacity and basicity, yielding a highly selective catalyst. Psathas et al. used asymmetrical-DN-FSP-deposited NiO or Pt
0 nanomaterials on the surface of Ta
2O
5 or the perovskite NaTaO
3 [72][53]. Single-step synthesis of the smallest produced NaTaO
3 (<15 nm), with finely dispersed NiO or Pt
0 (<3 nm). NaTaO
3/NiO produced from FSP had half the photocatalytic hydrogen production than those from DN-FSP. Also, DN-FSP had a ten times higher yield than the conventional deposition of wet-impregnated NiO. Similar results were found for the photocatalytic efficiency of NaTaO
3/Pt
0, which was 30% more photocatalytically active than the conventional liquid-Pt photo-deposition method
[54][36].