Direct numerical simulation of SpraySyn burner: Impact of liquid solvent

Abstract

The SpraySyn burner is a new system recently developed at the University of Duisburg-Essen to investigate experimentally nanoparticle synthesis in spray flames for a variety of materials. The current project aims at performing direct numerical simulations with detailed physicochemical models of configurations closely related to this burner. The effect of using different solvents to produce titanium-dioxide (TiO $_{2}$ ) nanoparticles is discussed in this work. The two solvents considered are o-xylene and ethanol mixed in liquid state with tetraisopropoxide to form TiO $_{2}$ . The liquid is injected into a pilot flame as dispersed spray with a carrier flow (dispersion gas). The resulting particle size distribution is examined as well. It is in particular observed that using ethanol leads to faster agglomeration and larger nanoparticles. This effect is qualitatively similar to that found when injecting smaller liquid spray droplets.

Keywords

Direct numerical simulation SpraySyn burner nanoparticle synthesis spray combustion

Introduction

In recent years, increasing attention has been paid to nanoparticle synthesis. This is because nanomaterials are found in many applications, such as color pigments in paints, carbon black in tires, drug delivery, cancer therapy, etc. Titanium dioxide is currently one of the most important nanoparticles because it is a natural oxide of the common element titanium, with negligible biological effects and low toxicity. Its classification as bio-inert material has opened the possibility to use it extensively in food products and as ingredient in a wide range of pharmaceuticals and cosmetics, such as toothpastes, sunscreens, etc. Consequently, many studies have been devoted to TiO $_{2}$ and other nanoparticles synthesis from combustion. This started with the formation of nanoparticle from gaseous flames.^1–4 Obviously, this conventional gas-phase process requires precursors that are either gaseous, or that can be vaporized and mixed with the burner gases before they react within the reaction chamber. Unfortunately, such precursors are only available for a limited number of elements and are often based on metal chlorides, metal organics, or organometallics that are very expensive and/or toxic.⁵ Hence, the focus is currently extended toward nanoparticle synthesis from spray flames,^6–14 allowing to overcome most of the problems listed above.

It has been increasingly recognized that close cooperation between experimentalists and modelers leads to high-quality databases and more accurate models. This collaboration often starts by organizing regular workshops discussing laboratory-scale experimental benchmarks, then asks the participants to extend the corresponding data set using their own experimental or numerical tools. At the end, the data from all participants are compared to understand the underlying processes. Examples of such workshops are (1) the international workshop on measurement and calculations of turbulent flames (TNFs),¹⁵ which is focusing on turbulent gaseous burners, (2) the workshops on turbulent combustion of spray (TCS), mostly addressing atmospheric and low-pressure spray burners,¹⁶ while (3) the engine combustion network (ECN) considers high pressure, dense sprays that are representative of engine conditions.¹⁷ These large projects investigate gaseous or spray flames, but without considering nanoparticle synthesis.

More recently, a collaborative research initiative was started in Germany (SPP1980, funded by the German Research Foundation—DFG—and entitled “Nanoparticle Synthesis in Spray Flames – SpraySyn: Measurement, Simulation, Processes”) to create a reference configuration, the “SpraySyn-burner” allowing to investigate in a systematic manner nanoparticle formation in spray flames. This burner shall deliver benchmark data for the corresponding research community.^5,18,19 One of the advantages of this burner is that it is designed from the start while taking into account the bottlenecks of companion numerical simulations.

Regarding computational fluid dynamics (CFD) simulations of TiO $_{2}$ nanoparticle generation in flames, Xu et al.²⁰ introduced a new approach for optimizing the process parameters and particle properties, based on the combined use of CFD with population balance-Monte Carlo (CFD-PBMC). In this work, Reynolds-averaged Navier-Stokes equations completed by the $k$ - $ϵ$ turbulence model and species transport-eddy dissipation model were solved to obtain flow fields and flame properties. The authors obtained good agreement with experimental data.

Torabmostaedi and Zhang²¹ developed and validated a CFD model to simulate the growth of TiO $_{2}$ nanoparticles in flame spray pyrolysis using coupled CFD-monodisperse population balance model (CFD-MPB) implemented within ANSYS-Fluent. The configuration is similar to that of the SpraySyn burner. The authors also report good agreement compared with experimental data.

Bringley et al.²² studied TiO $_{2}$ nanoparticle synthesis in a jet-wall stagnation flame in a 2D configuration using OpenFOAM to solve the gase-phase equations. Two different conditions were considered, either lean or stoichiometric. As expected, different combustion conditions lead to noticeably different particle size distributions.

For the SpraySyn burner configuration, most existing simulations rely on large eddy simulations, like, for instance, those documented by Schneider et al.,⁵ Rittler et al.,⁹ Prenting et al.,¹⁸ and Sellmann et al.²³ The present authors are involved as well in this collaborative DFG project, with the ultimate objective of providing direct numerical simulation (DNS) results for conditions as close as possible to those found in the real SpraySyn burner as it has been documented in our recent publications.^24,25 The latter publications can be considered as the first DNS of nanoparticle production in the SpraySyn burner.

In the present work, the production of titanium dioxide (TiO $_{2}$ ) nanoparticles from tetraisopropoxide (TTIP) as a liquid precursor, has been investigated using DNS of spray combustion under conditions closely related to that of the SpraySyn burner. Two different solvents are mixed separately with TTIP: o-xylene and ethanol. In two different simulations, the impact of these solvents is examined. Then, the diameter of the injected droplets has been changed to evaluate the impact of this property on nanoparticle synthesis.

The article is organized as follows: the “Introduction” section introduces the governing equations and models; numerical setups and configurations are discussed in the “Mathematical and numerical approaches” section; the DNS results are discussed in the “Results and discussion” section, before concluding in the “Conclusions” section.

Mathematical and numerical approaches

The simulations involve all three aggregate states: a main gaseous flow, liquid droplets, and solid nanoparticles. Therefore, three main numerical approaches are also employed: (1) (Eulerian) DNS to solve for the gas phase including detailed models for the corresponding chemical reactions; (2) a Lagrangian description for tracking spray droplets; and (3) a Eulerian approach to model the processes controlling the evolution of the nanoparticles (growth, aggregation, coagulation, etc.). Droplets, being noticeably smaller than the grid resolution and Kolmogorov scale, are modeled as point droplets with a variable diameter. All numerical models are integrated into the in-house DNS code called DINO, a Fortran90 code developed by our group during the last ten years. A sixth-order central finite-difference approach is used for the spatial discretization, while a semi-implicit third-order Runge-Kutta method with PyJac (analytical Jacobian matrix solver²⁶) are employed for temporal integration. The open-source library Cantera 2.4.0²⁷ is used to compute all chemical reactions, thermodynamic terms, and molecular transport processes in the gas phase. More details about DINO can be found in particular by Abdelsamie et al.,²⁸ Chi et al.,^29,30 and Abdelsamie et al.³¹ The Lagrangian model in DINO relies on the discrete particle simulation approach, DPS. Hence, the resulting simulations can be categorized as DNS-DPS. A two-way coupling between both phases (gas and spray) is implemented via the exchange of mass, momentum and energy. The droplet equations rely on the model first introduced by Abramzon and Sirignano,³² taking into account the modifications later suggested by Kitano et al.³³ All details concerning the implemented equations describing droplet movement as well as exchange of momentum, mass and heat between the liquid and the gas phases can be found in previous publications.^34,35

Concerning now the third modelling level, used to describe the evolution of the nanoparticles, the model developed by Kruis et al.³⁶ with improvements described by Weise et al.⁸ has been implemented. This model can be summarized as follows:

\frac{d N}{d t} = - \frac{1}{2} β N^{2} + I

(1)

\frac{d A}{d t} = - \frac{1}{τ_{s}} (A - N a_{s}) + I a_{0}

(2)

\frac{d V}{d t} = I v_{0}

(3)

In these equations,

N

is the nanoparticles’ concentration,

A

is the total surface area concentration,

V

is the total volume concentration,

v_{0}

is the monomer volume,

a_{0}

is the monomer surface area, and

I

is the nucleation rate. The coagulation frequency

β

is computed as

β = 4 π d_{c} D {[\frac{0.5 d_{c}}{d_{c} + g \sqrt{2}} + \frac{D \sqrt{2}}{0.5 c d_{c}}]}^{- 1}

(4)

g = \frac{[(d_{c} + L)^{3} - (d_{c}^{2} + L^{2})^{1.5}]}{3 L d_{c}} - d_{c}

(5)

L = \frac{8 D}{π c}

(6)

c = \sqrt{\frac{8 k_{b} T}{π ρ_{n} V}}

(7)

D = \frac{k_{b} T}{3 π μ d_{c}}

(8)

where

L

is the mean free path,

k_{b}

is the Boltzmann constant,

ρ_{n}

is the particle density,

T

is the gas temperature,

μ

is the viscosity of the gas,

c

is the particle velocity, and

D

is the diffusion coefficient of the particles. The primary diameter

d_{p}

, aggregate diameter

d_{a}

, and collision diameter

d_{c}

are computed as follows

d_{p} = (\frac{6 V}{A})

(9)

d_{a} = {(\frac{6 V}{π N})}^{1 / 3}

(10)

d_{c} = d_{p} n_{p}^{1 / 1.8}

(11)

n_{p} = {(\frac{d_{a}}{d_{p}})}^{3}

(12)

In equation (2), the surface area of the completely fused particles

a_{s} = {(\frac{V}{N v_{0}})}^{2 / 3} a_{0}

(13)

and the characteristic sintering time

τ_{s}

for titanium dioxide is computed similarly to Buesser et al.,³⁷

τ_{s} = 9.75 \times 10^{15} T d_{p}^{4} \exp (\frac{31000}{T})

(14)

Numerical configuration

As it has been discussed before, the main purpose of this work is to have conditions representative of the SpraySyn burner in the DNS. In the experiments, the main solvent is currently ethanol, which is mixed (in liquid state) with a precursor (iron(III)-nitrate nonahydrate) and then injected together with a dispersion gas (O $_{2}$ ) through a central injector. The liquid spray is then evaporated by the methane pilot flame. Here, the pilot flame entering the domain is composed of the reaction products obtained by burning a lean CH $_{4}$ /O $_{2}$ mixture (equivalence ratio of 0.65) after computing the corresponding equilibrium condition at constant enthalpy, as obtained from the Cantera package. The liquid solution (ethanol+precursor) is injected as a spray through the central tube of the burner, while the pilot flame and the protecting coflow enter through two annular regions surrounding the central injector. The complete description of the burner and experimental setup can be found bySchneider et al.⁵

Obviously, DNSs of the large and complex configuration shown in Figure 1 are extremely challenging. As a consequence, only the most relevant part of the domain has been considered in this first DNS study, as illustrated in Figure 1.

Figure 1.

(a) Schematic diagram showing the computational domain taken into account for the direct numerical simulations (DNS)s, projected onto the real burner and (b) typical instantaneous spray flame as obtained from the current DNS.

The dimensions of this DNS computational domain are 18 mm $\times$ 9 mm $\times$ 0.56 mm in streamwise, transverse, and crosswise directions, respectively. In the crosswise direction, only a small extension has been implemented, allowing the development of three-dimensional structures but reducing computational costs. The DNS domain is discretized over $16.7$ million grid points with 17.6 $μ$ m uniform grid spacing. Inflow/outflow boundary conditions are used in the streamwise direction, while periodic boundary conditions are considered in the other directions to simplify the process.

Injection conditions similar to that found in the experiment have been implemented: dispersion gas (O $_{2}$ ) is injected at a speed of $U_{j} = 91.34$ m/s, temperature of $T = 500$ K, and jet Reynolds number is 711. The pilot flame (CH $_{4}$ /air) enters the domain at a speed of $U_{p} = 3.71$ m/s; finally, the surrounding coflow (N $_{2}$ ) is injected at a speed of $U_{c} = 0.637$ m/s and temperature of $T = 500$ K. The pilot flame is injected at an equivalence ratio of $0.65$ , hence, its resulting peak flame temperature is $T \approx 1793$ K. It should be noted that no synthetic turbulence was injected through the boundary condition or initiated for these simulations.

In the simulations, initially monodisperse liquid droplets (containing solvent + TTIP), all starting with a diameter of $10 μ m$ and a temperature of $300$ K, are injected through the nozzle with a diameter of $d_{j} = 0.3$ mm. The pilot flame is injected through an annular area with inner and outer diameters of $1.2$ and $3$ mm, respectively. The coflow starts being injected beyond the radius of $3$ mm.

In the present simulations, two different solvents are tested: ethanol (currently used in the experiments) and o-xylene (which is expected to bring some advantages). Kinetics obviously play an important role for the final process outcome. In the current DNSs, skeletal kinetic mechanisms are used to describe both ethanol and o-xylene oxidations. For ethanol, it consists of 35 species and 87 elementary reactions. For o-xylene, it consists of 24 species and 107 elementary reactions.

These mechanisms are developed and optimized at the University of Duisburg-Essen based on the large systems published by Marinov³⁸ and Dagaut et al.³⁹ A mechanism describing the precursor really employed in the experiment (iron(III)-nitrate nonahydrate) is currently under development, but is not available yet. Therefore, in the present study, a mechanism for TTIP is used instead⁸; this is the best approximation at hand.

Results and discussion

In this section, the impact of using two different solvents to produce the same nanoparticle will be discussed. These solvents are o-xylene (Case I), and ethanol (Case II). In these two cases, the TTIP (precursor) is mixed with the solvent under the same conditions to produce ultimately TiO $_{2}$ considering the conditions mentioned in the previous section. Then, the impact of the initial liquid droplet diameter will be discussed.

The solvent is considered to be perfectly mixed in liquid state with the precursor. Then, this liquid mixture is injected through the central injector (Figure 1). It is assumed in the DNS that the spray droplets disperse immediately after injection, so that liquid jet breakup is not considered. The dispersed liquid droplets then start to evaporate due to the gaseous pilot flame. When reaching suitable conditions, the production of nanoparticles starts.

Impact of the solvent on nanoparticle synthesis

In order to describe the impact of the two different solvents, three sets of figures will be presented at four different time instants $t = 31 τ_{j}$ , $46 τ_{j}$ , $61 τ_{j}$ , and $76 τ_{j}$ , where $τ_{j} = d_{j} / U_{j}$ is the jet time-scale, equal to 3.28 $μ$ s here. Figure 2 shows the histogram of aggregate diameters of the generated nanoparticles, while Figures 3 and 4 present the scatter plots of TTIP mixture fraction in the gas mixture versus gas temperature and versus solvent mixture fraction in the gas mixture, respectively.

Figure 2.

Temporal evolution of normalized particle size distribution (PSD) for the nanoparticle aggregate diameter. From top to bottom, the time is $t = 31 τ_{j}$ , $46 τ_{j}$ , $61 τ_{j}$ , and $76 τ_{j}$ , respectively. (a) Case I (o-xylene) and (b) Case II (ethanol).

Figure 3.

Temporal evolution of scatter plot of $Y_{TTIP}$ versus gas temperature. From top to bottom, the time is $t = 31 τ_{j}$ , $46 τ_{j}$ , $61 τ_{j}$ , and $76 τ_{j}$ , respectively. (a) Case I (o-xylene) and (b) Case II (ethanol). TTIP: titanium tetraisopropoxide.

Figure 4.

Temporal evolution of scatter plot of $Y_{TTIP}$ versus mixture faction of the solvent in the gas mixture. From top to bottom, the time is $t = 31 τ_{j}$ , $46 τ_{j}$ , $61 τ_{j}$ , and $76 τ_{j}$ , respectively. (a) Case I (o-xylene) and (b) Case II (ethanol). TTIP: titanium tetraisopropoxide.

Looking at Figure 2, it can be observed that using ethanol as a solvent leads to larger nanoparticles at the same physical time (Figure 2(b)), compared to o-xylene (Figure 2(a)). This is a result of the differences in the thermodynamic, chemical, and transport properties of the solvents. One of the most important parameters controlling nanoparticle diameter is the evaporation rate of the liquid mixture; a faster evaporation would enable faster growth. The scatter plot showing the mass fraction of TTIP in the gas mixture versus gas temperature (Figure 3) shows that the evaporation rate with ethanol is higher than with o-xylene. This is already very clear at the first two time instants ( $t = 31 τ_{j}$ , and $46 τ_{j}$ ). At time $61 τ_{j}$ , the mixture fraction of TTIP in Case I peaks around $T = 1793$ K (Figure 3(a)), which corresponds to the interface between the spray jet and the pilot flame. A similar peak is observed at the same temperature in Case II (Figure 3(b)) but with a higher magnitude; simultaneously, TTIP appears also in the gas phase at noticeably higher temperatures. This means that the spray starts to ignite in case of ethanol (Case II) at $t = 61 τ_{j}$ , which is not seen in Case I. The observed differences become even more pronounced at later times (e.g. $t = 76 τ_{j}$ ).

Until this point, the relation between the distribution of mixture fraction of solvents and precursor in the gas mixture is not clear. For example, if this relation would be linear, it would become straightforward to predict the nanoparticle size distribution by knowing the behavior of the solvent. For o-xylene at early times ( $t = 31 τ_{j}$ , $46 τ_{j}$ , see Figure 4(a)), the quantities $Y_{TTIP}$ versus $Y_{Xylene}$ show a linear relation. When time goes on, vortical structures and turbulence start to play a noticeable role (even if the Reynolds numbers are still small in this set-up) and the behavior of precursor and solvent separate from each other in the gas mixture (see $t = 61 τ_{j}$ and $76 τ_{j}$ ). Looking at Figure 4(a), assuming a linear relation for Case I might be an acceptable assumption, particularly so at early times. On the other hand, for Case II (ethanol, see Figure 4(b)), the deviation starts very early and is much stronger, so that assuming a linear dependency would not hold. Apparently, the behavior of ethanol and TTIP differ very much in the low-turbulence gaseous environment of the SpraySyn burner. Since the density of the droplets containing o-xylene ( $ρ = 880$ kg/m $^{3}$ ) is noticeably larger than those containing ethanol ( $ρ = 789$ kg/m $^{3}$ ), the Stokes numbers and consequently the Lagrangian trajectories of the liquid droplets differ as well when injected into the same turbulent gaseous flow. This leads to different local nanoparticle concentrations $N$ as a function of the solvent. Figure 5 shows the nanoparticle concentration versus vorticity at two different time instants, $t = 61 τ_{j}$ and $76 τ_{j}$ . It is observed that, for all cases, the peak values of $N$ are found at places with low vorticity ( $ω_{x} \approx 0$ ). At the first time, only small differences are observed. Since o-xylene has a higher density than ethanol, it could be expected to find less droplets at higher vorticity with this solvent (stronger centrifugal effects), leading to a lower nanoparticle concentration at locations with higher vorticity. In the current work, the opposite behavior occurs, with higher nanoparticle concentration observed for o-xylene at high vorticity values compared to that of ethanol case. Where, all distributions of nanoparticle concentration show roughly a Gaussian shape with vorticity, but with a very broad tail for o-xylene (Figure 5(a)), which is not observed at all for ethanol (Figure 5(b)). This is because ethanol droplets ignite must faster than those containing o-xylene as solvent (see again Figure 3), leading to a more compact distribution.

Figure 5.

Temporal evolution of scatter plot of nanoparticle concentration $N$ versus x-component of vorticity in the gas mixture. From left to right, the time is $t = 61 τ_{j}$ , and $76 τ_{j}$ , respectively. (a) Case I (o-xylene) and (b) Case II (ethanol).

Impact of spray droplet size

In this section, the impact of the initial droplet size on nanoparticle formation is discussed. For this purpose, three cases with three different initial diameters have been computed by DNS: $d = 8 μ$ m, (b) $d = 10 μ$ m, and (c) $d = 12 μ$ m. The solvent used here is ethanol (currently employed in the experiments). The other conditions are the same as in previous section. The temporal evolution of the nanoparticle PSD is shown at three different times, $t = 46 τ_{j}$ , $61 τ_{j}$ , and $76 τ_{j}$ .

It is observed from Figure 6 that the PSD of the nanoparticles grows considerably faster when injecting spray droplets with a smaller initial diameter. This is due to the fact that smaller droplets for the same liquid mixture (solvent+precursor) lead to a much faster evaporation; thus, nanoparticle formation starts earlier. This can be confirmed by looking at the scatter plot of $Y_{TTIP}$ in the gas phase versus gas temperature, shown in Figure 7. The available TTIP in the gas mixture is noticeably larger for smaller spray droplets.

Figure 6.

Temporal evolution of normalized particle size distribution (PSD) for the nanoparticle aggregate diameter. From left to right, the time is $t = 46 τ_{j}$ , $61 τ_{j}$ , and $76 τ_{j}$ , respectively. The initial diameter of the spray droplets is (a) $d = 8 μ$ m, (b) $d = 10 μ$ m, and (c) $d = 12 μ$ m.

Figure 7.

Temporal evolution of scatter plot of $Y_{TTIP}$ in the gas phase versus gas temperature. From left to right, the time is $t = 46 τ_{j}$ , $61 τ_{j}$ , and $76 τ_{j}$ , respectively. The initial diameter of the spray droplets is (a) $d = 8 μ$ m, (b) $d = 10 μ$ m, and (c) $d = 12 μ$ m. TTIP: titanium tetraisopropoxide.

Conclusions

In this work, direct numerical simulations of a configuration similar to the SpraySyn burner have been conducted. The purpose of this burner is to produce nanoparticle materials from a spray flame in a controlled manner. For the presented results, two different solvents (o-xylene and ethanol) have been mixed with TTIP as a liquid precursor, with the final objective of producing titanium dioxide nanoparticles, TiO $_{2}$ . Liquid spray evaporation and nanoparticle synthesis are represented in the DNS code by three sets of coupled models. The impact of the solvents on the nanoparticle synthesis has been investigated. Then, the influence of the initial spray droplet diameter has been quantified.

It is observed that, when using ethanol as a solvent, the nanoparticles grow faster, leading to a larger mean nanoparticle diameter at the same time compared to o-xylene. For o-xylene, the mixture fraction of TTIP in the gas phase follows a quite linear relation with the mixture fraction of the evaporated solvent; this is not the case for ethanol. Due to differences in reactivity, the nanoparticles concentrate in low-vorticity regions when using ethanol as a solvent; while a much broader distribution is observed for o-xylene, showing a much stronger impact of flow structures and turbulence.

The initial diameter of the spray droplets has also a significant impact on nanoparticle production. Smaller droplets lead to the faster nanoparticle initiation and growth due to much faster evaporation and reduced inertia.

It is important to note that the present findings should be confirmed by considering a larger numerical domain and longer simulation times. This will be the subject of future publications.

Footnotes

Acknowledgements

The financial support of the Deutsche Forschungsgemeinschaft (DFG) for Abouelmagd Abdelsamie within the priority program SPP1980-SPRAYSYN “Nanoparticle Synthesis in Spray Flames SpraySyn: Measurement, Simulation, Processes” under Grant TH881/27-2 is gratefully acknowledged. The computer resources provided by the Jülich supercomputing center have been essential to obtain part of the results presented in this work.

Declaration of conflicting interests

The author(s) declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for this article’s research, authorship, and/or publication.

ORCID iD

Abouelmagd Abdelsamie

References

Kammler

Mädler

Pratsinis

. Flame synthesis of nanoparticles. Chem Eng Technol 2001; 24: 583–596.

Janzen

Knipping

Rellinghaus

, et al. Formation of silica-embedded iron-oxide nanoparticles in low-pressure flames. J Nanoparticle Res 2003; 5: 589–596.

Roth

. Particle synthesis in flame. Proc Combust Inst 2007; 32: 1773–1788.

Ren

Biswas

, et al. Flame aerosol synthesis of nanostructured materials and functional devices: Processing, modeling, and diagnostics. Prog Eng Combust Sci 2016; 55: 1–59.

Schneider

Suleiman

Menser

, et al. SpraySyn – a standardized burner configuration for nanoparticle synthesis in spray flames. Rev Sci Instrum 2019; 90: 085108.

Mädler

Kammler

Mueller

, et al. Controlled synthesis of nanostructured particles by flame spray pyrolysis. J Aeros Sci 2002; 33: 369–389.

Mueller

Mädler

Pratsinis

. Nanoparticle synthesis at high production rates by flame spray pyrolysis. Chem Eng Sci 2003; 58: 1969–1976.

Weise

Menser

Kaiser

, et al. Numerical investigation of the process steps in a spray flame reactor for nanoparticle synthesis. Proc Combust Inst 2015; 35: 2259–2266.

Rittler

Deng

Wlokas

, et al. Large eddy simulations of nanoparticle synthesis from flame spray pyrolysis. Proc Combust Inst 2017; 36: 1077–1087.

10.

Gao

Zhao

. Flame spray pyrolysis made Pt/TiO2 photocatalysts with ultralow platinum loading and high hydrogen production activity. Proceedings of the Combustion Institute 2021; 38: 6503–6511.

11.

Suleiman

Nanjaiah

Skenderovic

, et al. Atmospheric-pressure particle mass spectrometer for investigating particle growth in spray flames. J Aerosol Sci 2021; 158: 105827.

12.

Gonchikzhapov

Kasper

. Decomposition reactions of fe(co)5, fe(c5h5)2, and TTIP as precursors for the spray-flame synthesis of nanoparticles in partial spray evaporation at low temperatures. Industrial Engineering Chemistry Research 2020; 59: 8551–8561.

13.

Foo

Unterberger

Martins

FJWA

, et al. Investigating spray flames for nanoparticle synthesis via tomographic imaging using multi-simultaneous measurements (times) of emission. Opt Express 2022; 30: 15524–15545.

14.

Stodt

Liu

, et al. Phase-selective laser–induced breakdown spectroscopy in flame spray pyrolysis for iron oxide nanoparticle synthesis. Proceedings of the Combustion Institute 2021; 38: 1711–1718.

15.

TNF series; https://tnfworkshop.org/tnf15-workshop/. (1996-2022).

16.

TCS series; http://www.tcs-workshop.org/. (2009-2022).

17.

ECN series; https://ecn.sandia.gov/. (2011-2022).

18.

Prenting

Baik

Dreier

, et al. Liquid-phase temperature in the SpraySyn flame measured by two-color laser-induced fluorescence thermometry and simulated by LES. Proc Combust Inst 2022; 39(2): 2621–2630.

19.

SpraySyn Database, accessed 3 January 2022, available at; http://doi.org/10.17185/cenide/spraysyn/. (2022).

20.

Zhao

. CFD-population balance monte carlo simulation and numerical optimization for flame synthesis of TiO2 nanoparticles. Proceedings of the Combustion Institute 2017; 36: 1099–1108.

21.

Torabmostaedi

Zhang

. Numerical simulation of TiO2 nanoparticle synthesis by flame spray pyrolysis. Powder Technol 2018; 329: 426–433.

22.

Bringley

Manuputty

Lindberg

, et al. Simulations of TiO2 nanoparticles synthesised off-centreline in jet-wall stagnation flames. J Aerosol Sci 2022; 162: 105928.

23.

Sellmann

Wollny

Baik

, et al. Les of nanoparticle synthesis in the SpraySyn burner: a comparison against experiments. Powder Technol 2022; 404: 117466.

24.

Abdelsamie

Kruis

Wiggers

, et al. Nanoparticle formation and behavior in turbulent spray flames investigated by DNS. Flow Turbul Combust 2020; 105: 497–516.

25.

Abdelsamie

Chi

Nanjaiah

, et al. Direct numerical simulation of turbulent spray combustion in the SpraySyn burner: impact of injector geometry. Flow Turbul Combust 2021; 106: 453–469.

26.

Niemeyer

Curtis

Sung

. pyjac: analytical Jacobian generator for chemical kinetics. Comput Phys Commun 2017; 215: 188–203.

27.

Goodwin

. An open-source, extensible software suite for CVD process simulation. Chemical Vapor Deposition XVI and EUROCVD 2003; 14: 2003–2008.

28.

Abdelsamie

Fru

Oster

, et al. Towards direct numerical simulations of low-Mach number turbulent reacting and two-phase flows using immersed boundaries. Comput Fluids 2016; 131: 123–141.

29.

Chi

Janiga

Abdelsamie

, et al. DNS study of the optimal chemical markers for heat release in syngas flames. Flow Turbul Combust 2017; 98: 1117–1132.

30.

Chi

Abdelsamie

Thévenin

. Direct numerical simulations of hotspot-induced ignition in homogeneous hydrogen-air pre-mixtures and ignition spot tracking. Flow Turbul Combust 2018; 101: 103–121.

31.

Abdelsamie

Lartigue

Frouzakis

, et al. The Taylor–Green vortex as a benchmark for high-fidelity combustion simulations using low-Mach solvers. Comput Fluids 2021; 223: 104935.

32.

Abramzon

Sirignano

. Droplet vaporization model for spray combustion calculations. Int J Heat Mass Transfer 1989; 32: 1618–1618.

33.

Kitano

Nishio

Kurose

, et al. Effects of ambient pressure, gas temperature and combustion reaction on droplet evaporation. Combust Flame 2014; 161: 551–564.

34.

Abdelsamie

Thévenin

. Direct numerical simulation of spray evaporation and autoignition in a temporally-evolving jet. Proc Combust Inst 2017; 36: 2493–2502.

35.

Abdelsamie

Thévenin

. On the behavior of spray combustion in a turbulent spatially-evolving jet investigated by direct numerical simulation. Proc Combust Inst 2019; 37: 3373–3382.

36.

Kruis

Kusters

Pratsinis

, et al. A simple model for the evolution of the characteristics of aggregate particles undergoing coagulation and sintering. Aerosol Sci Tech 1993; 19: 514–526.

37.

Buesser

Gröhn

Pratsinis

. Sintering rate and mechanism of TiO2 nanoparticles by molecular dynamics. J Phys Chem C 2011; 115: 11030–11035.

38.

Marinov

. A detailed chemical kinetic model for high temperature ethanol oxidation. Int J Chem Kinet 1999; 31: 183–220.

39.

Dagaut

Ristori

Frassoldati

, et al. Experimental study of tetralin oxidation and kinetic modeling of its pyrolysis and oxidation. Energy Fuels 2013; 27: 1576–1585.