WO2010046595A1

WO2010046595A1 - Method and system for analysing the stoichiometry of particles and process for manufacturing particles with stoichiometry control

Info

Publication number: WO2010046595A1
Application number: PCT/FR2009/052008
Authority: WO
Inventors: Emmeric Frejafon; Tanguy Amodeo; Christophe Dutouquet
Original assignee: Institut National De L'environnement Industriel Et Des Risques
Priority date: 2008-10-21
Filing date: 2009-10-21
Publication date: 2010-04-29
Also published as: FR2937419B1; FR2937419A1

Abstract

Method of analysing composite particles in order to determine their stoichiometry, comprising the following steps: b1) by means of a laser shot (10), a plasma (12) is formed from an aerosol containing the particles to be analysed; b2) an emission spectrum (18) coming from the plasma (12) is acquired using an optical spectrometer (68) equipped with a detector (66, 70); and c) the stoichiometry of the particles is determined from the emission spectra (18) by calculating the proportion S or Sa/b between two elements a, b via the equation in which the factor N^a,I tot/N^b,Itot is determined by the equation. Analysis system for implementing this method. Advantageously, the stoichiometry is determined in virtually real time, enabling it to be used for controlling processes. Process for manufacturing particles by laser pyrolysis in which the stoichiometry is controlled using the above method of analysis.

Description

Method and system for analyzing particle stoichiometry; particle manufacturing process with stoichiometry control

The present invention relates to a method for analyzing composite particles having a diameter of less than 1 μm, making it possible to determine the stoichiometry thereof.

A composite particle is a particle of at least two elements (i.e., at least two elemental chemical species), and the particle stoichiometry refers to the relative proportions of the elements constituting the particles.

Usually, to determine the stoichiometry of such particles, a particle analysis is performed in the laboratory, using chemical analysis means such as inductively coupled plasma optical emission spectroscopy (ICP-AES). 'English' Inductively Coupled Plasma - Atomic Emission Spectroscopy ').

In practice, the measurement is carried out as follows: During manufacture of a batch of particles, at the end of manufacture, a representative sample of these particles is taken; it is then taken to an analytical laboratory, where it is analyzed. This method has the disadvantage of not being automatic, and not to be carried out online, that is to say directly on the production process or in the vicinity of it.

In addition, the results are not available in real time:

The analysis is deferred. As we have seen, they are obtained only after manufacture, and even worse, a certain time after the delivery to the laboratory operator of a sample of the particles studied. For this reason, while manufacturing is in progress, no measurement results are available: It can not be ensured that the particles are in accordance with the desired composition, let alone adjust the parameters of the process. As a result, in the event of a bad initial setting of the process parameters, a significant production loss is inevitable since an entire batch of nonconforming particles will have been manufactured.

A first objective of the invention is to propose a method for analyzing composite particles with a diameter of less than 1 μm, making it possible to determine the stoichiometry thereof, a method in which, at least once, steps b1 are executed. and b2): bl) laser firing (10) on a stream of an aerosol containing the particles to be analyzed, so as to form a plasma (12), and b2) acquiring at least one emission spectrum (18) from the plasma using a light collection device (66) connected to an optical spectrometer (68) equipped with a detector (70), the acquisition taking place between an initial instant TD and a final instant T _F , these two moments being counted from the moment of the laser shot; and then, c) from the emission spectrum (s) (18) obtained, the stoichiometry of the particles is determined; a method that is automatic, that can provide reliable, reproducible and substantially real-time results, so that it can be used on the production line.

This objective is achieved thanks to the fact that in this method, step c) comprises the following operation:

the proportion between two elements a, b, by the equation (Sl) is calculated:

ς _ _ ^{t to} early early _

^~ N ^{b ~} N ^bJ -4- n ^b ^{~ l ~}

in this, the factor N ^ _t ot / N ¹³ ' ¹ _tot is determined by the equation (S2):

in which ; a, b are the two elements whose ratio S _a / t _{> is determined} , ij is the transition considered for the element a, between the levels i and j,

N ^a τot is the total number of atoms of the element a, the sum of the atoms N ^aI τo _t in the neutral state I, ie the degree of ionization 0, and of the atoms N ^a ' ^π τo _t at the first ionized state II, i.e. at the degree of ionization +1, I ^{a / I} ij is the total intensity corresponding to a line of the element a in the neutral state, corresponding to the transition between states i and j, h ^a '\ _j is the wavelength of radiation emitted during this transition, ^ά' ¹ is the partition function of the element in the neutral state. g, and g _k are the statistical weights respectively of level i of element a and level k of element b,

At ³ ' ¹ , -, is the transition probability between the levels i and j for the element a in the neutral state, E ^a , is the energy of the higher level i for the element a in the neutral state , k is Boltzmann's constant,

T _e is the electronic temperature.

As can be seen, this method of analysis is a method of the "LIBS" (Laser Induced Breakdown Spectroscopy) type or laser-induced plasma emission spectroscopy.

The particles are measured in suspension in a gas, and form with it an aerosol. Of course, in order not to prevent or disturb the measurement, the gas or gases constituting mainly the aerosol does not contain any element also contained by the particles. Advantageously, the measurement is performed on a stream of this aerosol, that is to say on a moving aerosol, for example in a pipe or equivalent.

The advantage of measuring the particles when they are diluted in a gas, in the form of an aerosol, is that it is very often the form in which they are produced. In this way, the analysis method according to the invention can be directly carried out in the aerosol stream containing the particles, without any prior treatment. The aerosol is simply taken from the aerosol stream, for example by means of a bypass line disposed in a branch line on an exit line of an aerosol generating reactor.

Advantageously, the measurement on a flow (here, an aerosol flow) allows a continuous measurement, and in particular an automatic measurement. The particles may be in the liquid phase and / or solid. As can be seen, the different steps of the analysis method can be performed automatically, whereby the results can be obtained very quickly (the acquisition of a spectrum of emission, as well as the data processing, can be made almost immediately, using a spectrometer equipped with a detector and a computer type PC). Despite the low density of the particles contained in the aerosol (in number of particles per cm ³ ), especially with respect to a solid body, it appears that the emission spectrum from the particles present in the aerosol is of sufficient quality to allow its acquisition by a spectrometer and to enable the determination of the stoichiometry of the particles present effectively. Most often, the initial time T ₀ corresponds to the beginning of the integration of the emission spectrum by the detector associated with the optical spectrometer; and the integration continues continuously until the final moment TF. Thus the interval of TD at T _{F is} the exposure time, or integration time, of the detector, used to obtain a transmission spectrum.

That being so, not all emission spectra are necessarily acquired during the whole of this interval; some spectra may be limited to a defined portion of this range to improve the signal-to-noise ratio for some spectral lines.

The use of the equations (Sl) and (S2) supposes that a single temperature governs the evolution of the population of atoms.

From the recorded emission spectra or spectra, the formulas (Sl) and (S2) make it possible to calculate the stoichiometry of the particles. In general, the equation (Sl) can be used in its general form, which takes into account the numbers of atoms of the elements considered under all of their degrees of ionization. In the formula indicated, on the other hand, only the degrees of ionization 0 and 1 have been taken into account, that is to say the atoms in the neutral state I, or in the first ionized state IL This choice is made because in In general, the quantities of atoms present at higher degrees of ionization (2 and more) are negligible in comparison with those present at degrees of ionization 0 and 1, at least under measurement conditions close to those described herein. document. To calculate the value of the term (1 + N ³ ' ¹ _t o _t / N' ^a ' ^π _tot ) / (1 + N ^W _tot / N ^AI1 _tot ) of the equation (S2), different ways of doing things are possible :

In one implementation mode, in step c) of the method of analysis, the factor N ³ ^'11 _t o _t / N ^3' ¹ _earlier in Equation (Sl) is determined by the equation ( S3):

where m _e is the mass of the electron, h is the Planck constant, and

Eion is the ionization energy of an atom of the element a passing from the neutral state I (atom in its ground state) to the state once ionized IL

The advantage of this approach is that for each element, several lines, not just one, can be used simultaneously to determine the proportion of this element relative to the others in the particles (provided that the element generates several lines exploitable in the spectrum (s) of emission). This can increase the accuracy of the measurement.

It will further be understood that in a general manner, in the case where the elements present in the plasma are also taken into account at one or more degrees of ionization greater than 1 (for example, 2, 3, etc.), the equation (S3) can be generalized, and we then use equation (S3) in a more general form, in which the indices 1,11 are replaced by pl, p, respectively, which gives the ratio in number of elements present in two consecutive degrees of ionization p-1 and p.

According to another implementation mode, the factor N ^has' _t ot ^π ^/ N ^{/ I} _soon equation (Sl) can be understood as follows: we choose a time T ₀ such that from At the instant T ₀ , the ionized forms of the elements constituting the particles are substantially in number at least ten times smaller than the number of the same elements in the neutral state. Thanks to this, in step c), the factors

"Of the equation (Sl) are considered negligible. Thus, the proportion S _a / b between two elements a, b, according to (Sl), satisfies:

Sa / b = N ^has _t o _t / NV = N ^a a ', ¹ I _tot / N, b, I early

In the particular case of particles comprising only two elements, the stoichiometric is defined by the ratio of the amounts of these two components. Also, the stoichiometry is given by a single equation directly giving the ratio of the total numbers of atoms of the two species, since the total numbers of atoms of an element are then substantially equal to the number of atoms in the neutral state.

Taking into account the above hypothesis in the equations, it follows that the stoichiometry is given by (Sl-2);

The method of analysis according to the invention can be refined by resorting to one or more of the following improvements:

According to one embodiment, step c) comprises the following step: the stoichiometry is determined by choosing as value for a relative proportion between two elements, the value for which is minimal a difference between the observed spectrum, and the spectrum; simulated by a plasma simulator.

According to one embodiment, the sequence of laser firing steps, or step b1) and acquisition of at least one spectrum, or step b2) is repeated at least a hundred times, and step c) is performed by operating accumulated information on all emission spectra thus acquired. The accumulation of the measurement results makes it possible to reduce the dispersion of the results obtained. The exploitation of the information accumulated over all the acquired emission spectra can be achieved by determining an emission spectrum representative of the series of acquired acquisition spectra, for example by averaging them, adding them together, or using any other analogous method to take into account the information contained in the different emission spectra, to synthesize it into a single spectrum.

According to one embodiment, the plasma is optically thin between the TD and TF instants. Thus, the acquisition of the emission spectrum (s) necessarily occurs while the plasma is optically thin, and the spectrum (s) is or are not degraded due to the plasma reabsorption of certain spectral lines. According to one embodiment, the plasma is at the local thermodynamic equilibrium between the instants T ₀ and TF. Thus, advantageously the acquisition of the emission spectrum (s) is done while the plasma is at the local thermodynamic equilibrium, which is a necessary condition for the validity of the equation (S2). A partial thermodynamic equilibrium condition may, however, be sufficient in some cases.

According to one embodiment, in step c), each of the spectra used to determine the relative proportions of the different elements constituting the particles is acquired continuously from time T ₀ and up to time T _F. In other words, the relative proportions of all the components are obtained from emission spectra which are all acquired from time Td to time Tf. Thanks to this, the implementation of the method remains simple. According to one embodiment, the aerosol comprises at least two types of composite particles.

According to one embodiment, the aerosol is helium-free. The method of analysis does not require, or at least not necessarily, the injection of helium to slow the recombination of atoms in the plasma. Also, the analysis method remains simple and can be applied directly to many particle production aerosols, in particular nanoparticles such as silicon carbides, produced in an argon stream by laser pyrolysis.

According to one embodiment, the plasma generated during a step b1) contains at least 300 particles, and preferably at least 300 particles.

1000 particles, whose radiation after vaporization is acquired in step b2) to allow the determination of the particle stoichiometry in step c). This result can be achieved by giving appropriate values for the concentration of particles in the flux, the intensity of the laser and the focusing thereof towards the plasma formation point. With such a large number of particles, it avoids threshold effects that could occur with a smaller number of particles, and could lead to an error in the determined stoichiometric value. According to one embodiment, in step b1), the aerosol in which the plasma is generated contains at least ¹⁰ particles per cubic centimeter, the particle size being between 10 nm and 100 nm. Indeed, a relatively high particle density in the aerosol promotes high measurement accuracy.

According to one embodiment, before step b1), a step a) takes place, during which the concentration of the particles in the aerosol is adjusted upwards or downwards. This avoids two possible pitfalls: - If the particle density is too low, the signal-to-noise ratio of the emission spectra drops, because the number of particles present in the plasma decreases thus reducing the intensity of the signal, and inducing variations in the results unrelated to the stoichiometry of the components in the particles; - If the density of particles is too large, the plasma is no longer optically thin, there is reabsorption of the photons emitted, and the results obtained by the method lose precision.

According to one embodiment, the analysis method is used to determine the stoichiometry of particles with a maximum diameter of less than 100 nm.

The method has thus been used, and proved effective, for the measurement of the stoichiometry of silicon carbide particles, whose diameters are between 30 nm and 100 nm. An example of such measures is presented in the following. These particles may in particular be particles of silicon carbides, titanium or boron; titanium oxides; or alternatively iron oxide Fβ ₂ θ ₃ .

In the case of particles with oxygen, care should be taken to eliminate any risk of explosion in the plasma by diluting the particles in an inert gas. According to one embodiment, in step b1), in the aerosol, the particles are suspended in a monoatomic inert gas such as argon, helium or neon. The method according to the invention can however more generally be implemented in other gases. In the latter case, specific precautions may be necessary if necessary (for example, in air), to take into account the possible explosiveness, and the formation of molecular species in the plasma.

A second object of the invention is to define a process for manufacturing composite particles with a diameter of less than 1 μm per laser pyrolysis, a process that can be adjusted or optimized during manufacture, thanks to reliable information and obtained substantially in real-time indicating the stoichiometry of the particles produced.

This objective is achieved by the fact that the method comprises the on-line control of the stoichiometry of the particles produced, by means of the analysis method defined above.

Preferably, the control is carried out immediately downstream of the point of pyrolysis on the production line, before any transformation of the particles. By 'online control' is meant here a control carried out on the production line itself or in the vicinity thereof. Advantageously, the analysis method presented above is extremely fast, which allows to benefit from the results almost immediately, and to exploit them to adjust or optimize the manufacturing process without waiting for the end of the current batch.

A third object of the invention is to define an analysis system for determining the stoichiometry of composite particles having a diameter of less than 1 μm, the system comprising:

a chamber interposed on a channel, the channel being able to convey an aerosol containing the particles to be analyzed;

a laser capable of firing on the aerosol as it passes through the chamber to locally form a plasma;

a device for collecting the light emitted by the plasma;

an optical spectrometer equipped with a detector towards which the collected light is directed;

calculating means able, from the emission spectrum or spectra obtained, to determine the particle stoichiometry; system that can provide results automatically, quickly, in order to provide fast and reproducible results, especially for implementation on production line.

This objective is achieved thanks to the fact that the calculation means (76) are designed to calculate the stoichiometry as follows: the proportion S or S _a / _b between two elements a, b is calculated using the equation (Sl):

in which the factor N _{t t} o _t / N ¹³ ' ¹ _early is determined by equation (S2):

where: a, b are the two elements whose ratio S _a / t _{> /} ij is the transition considered for the element a, between the levels i and j, N ^a χo _t is the total number of atoms of the element a, sum of the atoms N ^aI τo _t in the neutral state I, that is to say the degree of ionization 0, and of the atoms N ^a ' ^π τo _t at the first ionized state II it is ie the degree of ionization +1, p ¹ '^ is the total intensity corresponding to a neutral line of element a, corresponding to the transition between states i and j, λ ^{a / I} i _j is the wavelength of the radiation emitted during this transition, Z ³ ' ¹ is the partition function of the element a in the neutral state, g _f and gk are the statistical weights respectively of the level i of the element a and the level k of the element b,

At ⁸⁴ _Jj is the probability of transition between the levels i and j for the element a, in the neutral state

E ^a 'i is the energy of the higher level i for the element a in the neutral state, k is the Boltzmann constant, T _e is the electronic temperature.

The light detector associated with the spectrometer is generally an intensified CCD camera (or ICCD).

According to one embodiment, in the analysis system, the calculation means determine the factor N ^a ' ^π _t o _t / N ^{a / I} t _ôt of the equation (Sl) by using the equation (S3):

where m _e is the mass of the electron, h is the Planck constant, and

E _10n is the ionization energy of an atom of the element a passing from the neutral state I to the state once ionized IL

According to one embodiment, the light collection means are provided for collecting light only from the initial instant TD, from which the ionized forms of the elements constituting the particles are substantially in number at least ten times less than the number of the same elements in the neutral state, whereby in step c), the factors N ³⁴ WN ³ ' ¹ early and N ¹³⁴ WN "' ¹ early in equation (Sl) are considered negligible , and thus for the proportion Sa / b between two elements a, b, according to (Sl), verify:

According to one embodiment, in the analysis system the channel is installed between two points of an aerosol transport pipe, a pressure difference between these two points inducing the movement of the aerosol in the channel. Thus, advantageously the aerosol circulation in the channel is spontaneously under the effect of the pressure difference, and it is not necessary to provide means for moving the aerosol, such as for example a specific pump. .

According to one embodiment, the analysis system further comprises a gas supply channel, for adding gas into the aerosol upstream of the chamber to reduce the concentration of particles in the aerosol.

The invention will be better understood and its advantages will appear better on reading the detailed description which follows, of embodiments shown by way of non-limiting examples. The description refers to the accompanying drawings, in which: FIG. 1 is a schematic view showing the method of analysis by Laser Induced Plasma Emission Spectroscopy (LIBS);

FIG. 2 is a diagram showing an emission spectrum obtained on a plasma of silicon carbide particles; - Figures 3 and 4 are schematic views showing an analysis system according to the invention, for determining the stoichiometry of composite particles, installed on a production line of these particles, respectively in top view and in perspective; FIG. 5 is a diagram showing an emission spectrum resulting from measurements made, as compared with a calculated spectrum simulating two argon lines, for calculating the electron density;

FIG. 6 is a diagram showing the variation curves of the electronic density as a function of time, for the SiC and SiCs particles _; FIG. 7 is a Boltzmann diagram used for the determination of the electronic temperature;

FIG. 8 is a diagram showing the evolution curve of the electronic temperature T _e as a function of time during the manufacture of silicon carbide SiC particles; FIG. 9 is an evolution diagram of the intensity I of a silicon line as a function of the concentration C of the silicon; and

FIG. 10 is a diagram showing the impact of the variations of the electronic temperature T _θ

IMPLEMENTATION OF THE SIC PARTICLE PROCESS _x

Preamble: the LIBS method

First, the steps of the LIBS measurement method are recalled in relation to Figure 1. The LIBS measurement method is a basic analysis method, in two stages; At a time T ₀ , a short laser pulse 10 is sent, or laser shot, whose duration varies from a few femtoseconds to a few nanoseconds, to the surface of solids, liquids or even in an aerosol, which is to be determined the composition. A small amount of material is then heated in the state of ionized gas thus forming a plasma 12, At the heart of this plasma generated by the intense laser pulse sent, the material is vaporized, the molecules are dissociated. From a time T ₀ , subsequent to T ₀ , the radiation emitted by the plasma is acquired by means of an emitted light collection device 14, connected to an optical spectrometer 16 equipped with a detector, such as an ICCD camera; the composition of the sample studied is then determined using calculation means (computer 20), using the spectrum (s) 18 emitted by the plasma atoms 12.

FIG. 2 represents an optical emission spectrum obtained during the vaporization of silicon carbide nanoparticles (SiC _x ) in a plasma created by laser, spectrum obtained by means of an analysis system as presented previously. This spectrum constitutes the analytical signal or LIBS signal, and comprises different lines whose wavelength λ are characteristic of the elements present in the plasma, and make it possible to identify the latter. In the case presented, the spectrum comprises a line 22 characteristic of carbon C and a set of lines 24 characteristic of silicon Si.

Referring to Figures 3 and 4, an embodiment of the invention will now be presented.

The analytical method presented serves to determine the stoichiometry of silicon carbide particles (SiC _x ) produced by a production line 26 by laser pyrolysis. On this production line 26 is installed an analysis system 28 according to the invention, for determining the stoichiometry of the particles produced.

The production line 26 comprises a reactor 30 in which the nanoparticles are synthesized, shown in FIG. 3 (and schematized in a very simplified manner in FIG. 4). Gas injection means (not shown) make it possible to maintain this reactor 30 under an argon atmosphere.

The reactive gases, SiH ₄ and CzH ₂ , necessary for the production of SiC _x nanoparticles, are conducted into the reactor via a pipe 32 _# and injected through a nozzle 34 located in the center of the reactor 30 in its lower part. A high-power continuous laser beam 36, a few centimeters wide produced by a CO ₂ laser 37 passes through the reactor 30 perpendicularly to the axis of the nozzle 34. The beam is focused by a lens 38 in the direction perpendicular to its axis. of propagation, precisely at the point of injection P of the reactive gases, just in These gases absorb energy from the laser and dissociate before recombining in the pyrolysis flame to form SiC _x silicon carbide nanoparticles. These nanoparticles are then removed from the reactor in a conduit 40 by the injected argon flow, used as a carrier gas, the argon flow mixed with the nanoparticles constituting the 'product stream'.

The product stream passes through the conduit 40 to a collector 42, in which the nanoparticles produced are stored. The remainder of the flow produced, that is to say the gases contained therein, is sucked by a pump 44 placed downstream of the collector (FIG. 4) and connected to the latter by a suction duct 43.

The _erme analysis system 28 per _year ^ t ^ _e ^ j _erm _dinner | _has stoichiometry of SiC _x nanoparticles produced will now be presented. A sampling channel 46 placed in branch on the conduit 40 immediately downstream of the reactor 30 in which the pyrolise takes place, takes a "withdrawn flow" which is a fraction of the product stream. This sampling channel 46 consists of a hollow metal rod with an inside diameter of a few millimeters. Its upstream end (sampling end) is bevelled at an angle of about 55 to 60 degrees with respect to the straight section of the duct 40.

The stream taken is conveyed via the sampling channel 46 to a LIBS analysis cell 50, in which it will be subjected to analysis. The sampling channel 46 is arranged with radii of curvature such that the collisions of nanoparticles against the walls of the channel is minimized.

A first valve 48 is installed on the channel 46 upstream of the LIBS cell 50 in order to be able to stop the admission of the flow taken from the cell. The output of the LIBS analysis cell 50 is connected to a vacuum vial 52 via a downstream portion of the sampling channel 46. In the vacuum vial 52, a fraction of the nanoparticles of the sampled flow is collected for delayed chemical analyzes. The remainder of the sampled flow is evacuated using a flexible hose 54 connecting the outlet of the vacuum flask to the exhaust duct 43 downstream of the collector 42 of nanoparticles. The pressure difference existing between the connection point B of the sampling channel 46 on the production line 40, and the connection point C of the flexible hose 54 connected to the suction duct 43 downstream of the collector 42, is sufficient to allow the flow of the stream taken through the cell 50 to determine the stoichiometry of the particles.

A filter 55 and a second valve 56 are arranged on the flexible pipe 54 in order respectively to remove any remaining nanoparticles in suspension in the sampled flow, and to be able to stop the circulation of the sampled flow. Thus, closing the two valves 48 and 56 makes it possible to isolate the LIBS analysis cell 50 during maintenance operations or configuration changes in the production line by laser pyrolysis. The LIBS analysis cell 50 is provided with three openings for the optical sights. The first opening 58 is equipped with a convex fused silica piano lens with a focal length of 50 mm by means of which laser pulses of 5 ns duration, wavelength 1064 nm, energy 50 mJ delivered by a 60 Nd: YAG laser operating with a firing frequency of 20 Hz are focused in the internal chamber of the cell 50. A plasma is thus created in the sampled stream, consisting of a mixture of argon and nanoparticles. . A porthole 62 in quartz, whose input face is disposed perpendicularly to the faces of the two other portholes (58, 64), allows the passage of the light emitted by the plasma outside the cell. This radiation (or LIBS signal), containing the information on the chemical composition and the stoichiometry, is collected using a telescope 66, whose line of sight is placed perpendicularly to the laser beam. The telescope 66 is connected to a spectrometer 68 surmounted by an intensified CCD camera 70 with the aid of an optical fiber 72 having a core diameter of 50 μm. The third port 64 is installed facing the focusing lens to allow the escape of the beam which is then stopped by means of a beam stop device 74.

The data received by the spectrometer 68 equipped with the detector 70, that is to say the received emission spectra, are transmitted to a computer or computer 76 for determining the stoichiometry of the particles. DETERMINATION OF STOICHIOMETRY Recording of emission spectra

To determine the stoichiometry by implementing the analysis method according to the invention, using the analysis system 28, the procedure is as follows. During the production of nanoparticles, the valves 48 and 56 are opened so as to allow the passage of a stream taken through the LIBS analysis cell 50.

We send a series of laser shots of short duration on the flow taken. As a result of each laser firing, the generated emission spectrum is recorded after a TD delay with respect to the laser firing, during an integration time Δt (or exposure time), thus going up to a time T _F such that T _F = TD + Δt.

Plasma induced by laser firing on a flux derived from a manufacture of silicon carbide particles (SiC _x ) has a lifetime of about one hundred microseconds. The optical detection equipment (telescope 66, camera 70, spectrometer 68) makes it possible to perform a time-resolved plasma analysis. By varying the integration period Δt (interval between TD and TF) this equipment can provide optical emission spectra corresponding to different instants of the evolution of the plasma, in order to optimize the values of TD and T _F.

Each recorded spectrum corresponding to a single plasma, consequence of a single laser shot, is in principle sufficient for a qualitative analysis or analysis of the chemical composition, but is too little intense to be exploited quantitatively because of signal / noise ratios or unfavorable signal / continuum. The recordings of the spectra are thus accumulated on several laser shots. VERIFICATION OF THE CONDITIONS OF VALIDITY OF (S2) Before implementing the analysis system 28 which has just been presented, in order to proceed to the determination of the stoichiometry of batches of particles produced, it is necessary to verify that the conditions of applicability of the method of analysis are well fulfilled. These checks also make it possible to choose certain parameters of the method, such as the time window during which the emission spectrum is acquired.

The following operations and checks are carried out:

The electron density and the electronic temperature or plasma temperature are calculated;

The period during which the plasma is at the local thermodynamic equilibrium (ETL), during which it must satisfy one of the criteria of the local thermodynamic equilibrium, is then determined; The period during which the density of ions is low relative to the density of neutrals is determined for each element by studying the emission spectra;

It is verified that the plasma is optically thin for the lines of the elements composing the nanoparticles; and

According to the three preceding points, a suitable time window is chosen for the recording of the spectra.

These different verifications are detailed below. A. PRELIMINARY CALCULATIONS Electronic density

In a first step, the electron density is determined by measuring the Stark broadening, to which the density is in fact directly proportional according to the following formula; (S4) Δλstark = 2 W _r ef * N _e / N ₆ ^ref in which:

Δλs _t ark is the measured Stark width (FWHM)

W _ref is the half-width for the half-maximum of the Stark parameter for a given electron density N _e ^ref . The widening of a line may be due to several effects. In the example presented, it is the Stark effect that dominates. The width of the line is thus the superposition, or convolution, of the width inherent to the instrumentation (or width of apparatus) with the width due to the Stark effect. The enlargement caused by the instrumentation is measured using a mercury-argon spectral calibration lamp, which allows the application of a preliminary correction to the emission spectra.

Measuring then Δλ of a line from the emission spectrum, N _{6 is} then directly deduced, the values W _ref and N _e ^ref being tabulated;

The electron density is calculated for different spectral lines in order to increase the reliability of the value used.

In addition, the value of the electron density is determined for a series of time-phased emission spectra so as to determine the variation of the electron density N ₆ in the plasma as a function of time. The validity of the results obtained is confirmed by comparing the measurements made with a theoretical spectrum.

FIG. 5 thus presents an emission spectrum obtained with a delay T ₀ of 4.2 μs and an integration time Δt (interval between T ₀ and TF) of 2 μs (curve 78), illustrated with respect to a calculated spectrum (curve 79).

The emission spectrum is essentially limited to the two lines of Argon at 750.386 nm and 751.465 nm respectively.

On the ordinate is carried the intensity of the radiation (represented with an arbitrary unit), while on the abscissa is carried the wavelength of the radiation in nanometers.

It is found that for an electron density of 9.10 ¹⁶ cm ^"3, the calculated spectrum and the measured spectrum are substantially identical.

FIG. 6 shows the variation curves of the electronic density N _e as a function of time, in the cases where the particles studied are respectively SiC and SiCg.

In particular, it can be seen that the electronic densities are similar for these two elements, at least for a wide time range between 2 μs and 10 μs. During the first moments after the laser firing (t <2 μs), the measurement accuracy of the electron density N _e is however rather low, because of the strong Stark broadening effect during this phase.

Electronic temperature

In a second step, after the electronic density, the electronic temperature is determined.

This is calculated conventionally using a Boltzmann diagram.

One of the Boltzmann diagrams used, shown in Figure 7, was used to calculate the plasma temperature. This temperature is calculated by taking the logarithm of the equation of the intensity of a line, according to the formula:

(S5) In (I ₀ A ₀ / g, A ₀ ) = - E ₁ / kT _e + C in which

I ₀ is the intensity of the spectral line for the transition between energy levels i and j; λ, _] is the wavelength of the radiation emitted during this transition; g, is is the statistical weight of the energy level i, A ₀ is the transition probability between the levels i and j, E ₁ is the energy of the higher level ik is the Boltzmann constant, T _e is the electronic temperature ,

It appears immediately on reading equation (S5) that the temperature is deduced immediately from the slope of the curve.

This curve is given in FIG. 7, which shows the evolution of a logarithmic function, including the electronic temperature T _e of the plasma as a function of the E energies of the higher levels of the transitions, expressed in electron volts. The y-axis is a logarithm without dimension.

For this calculation, seventeen lines of argon were used, in order to reduce the uncertainty on the value found. The measurement is carried out for SiC nanoparticles. The measured electronic temperature is 14870 +/- 1400 K.

As for the electronic density N _e , the evolution of the electronic temperature T _e is determined as a function of time.

FIG. 8 shows the evolution curve of the electronic temperature Te (in ordinate, in degrees K) as a function of time T ₀ (in abscissa, in μs) after which emission spectra are acquired, during the fabrication of SiC particles.

In addition, it has been observed that the electronic temperature depends little on stoichiometry and concentrations of the species.

Criterion of validity of the local thermodynamic equilibrium

In a third step, the temperature and the electron density having been determined as a function of time, the time window is determined for which the plasma is at the local thermodynamic equilibrium. For this, the criterion of Mac Whirter is used (other similar criteria can be used):

(S6) N _e > N _L = 1.6. 10 ¹² (T _e ) ^1/2 . (ΔE) ³ where ΔE is the highest energy difference between high and low populated energy states, between which a transition is possible, the plasma being at equilibrium local thermodynamics. In the calculation, the electronic temperature chosen is that obtained after a delay T ₀ of 10 μs, that is to say the maximum time during which the electronic density could be determined; this value being approximately 12300 K according to FIG. 8. The value retained for ΔE is the energy difference for the two levels generating the argon line at 419.10 nm; this energy difference is then worth ΔE = 2.95 eV. From these values of T ₆ and .DELTA.E, N is obtained 4.5.10 _L = ¹⁵ cm ^"3.

From the diagram giving the electron density (figure 6), the hypothesis according to which the plasma is at the local thermodynamic equilibrium (notament for T ₀ <10 μs) is thus well verified.

Identification of the spectral lines and evaluation of the ratio of the density of ions on the density of neutrals In a fourth time, the qualitative study of the spectra resulting from the analysis of the plasma resolved in time makes it possible to evaluate the fraction of ions present in the plasma as a function of time.

The elements in the ionized state are considered to have disappeared when their characteristic lines no longer appear on the emission spectra. In fact, these elements have not entirely disappeared, but become rare is because of ion recombination process, so that their number is more than l / 10 ^th or l / 100 ^th of the items in the neutral state.

It has been possible to verify by study of the emission spectra obtained, that this condition is satisfied when the delay T ₀ is greater than 2 μs.

Optically thin plasma

In a fifth step, it is verified that the plasma is optically thin. A first verification made it possible to verify on the emission spectra that they do not have autorenversed lines, indicating a strong self-absorption.

In addition, a series of measurements have been carried out on aerosols having different concentrations of silicon. These tests made it possible to trace the evolution curve of the intensity I of the silicon line at 288.157 nm as a function of the concentration C of the silicon in the aerosol, in g / m 3, presented in FIG. 9. Figure 9 shows the measurements taken, represented by small circles, and a linear regression line passing through the origin and calculated to be representative of these measurements.

It can be seen that the intensity of the emission line is proportional to the concentration of silicon in the concentration range studied, which has led to the conclusion that there is almost no autoabsorption under the operating conditions. .

In addition, for carbon, its self-absorption rate may be considered negligible, since its transition probability is lower than that of silicon at 288.157 nm.

Finally, it has also been verified that there is no saturation effect of the emission spectrum due to incomplete vaporization of the particles in the laser-induced plasma. This check was made as follows: The particles produced by laser pyrolysis are of a diameter between 20 and 100 nm. A series of measurements was carried out and showed a linear relationship between the size of the particles produced and the intensity of the emission spectrum during the implementation of the method according to the invention. It has been concluded that there is no saturation effect of the emission spectrum due to incomplete or unrealized vaporization of larger diameter particles, and thus, for these particle sizes, the plasma is optically thin.

Choice of the time window In a sixth time, the adequate temporal window to obtain significant results has been determined. This time window is the interval extending between the two parameters T ₀ , the time delay between the laser firing and the start of acquisition of the emission spectrum; and T _F , equal to TD + Δt, where Δt is the integration time of the intensified CCD camera 70 associated with the spectrometer 68.

This time window is chosen according to three criteria: First, it must be included in the period during which the plasma is at the local thermodynamic equilibrium. The maximum duration is therefore 10 μs, the duration beyond which the electronic density could no longer be measured and therefore be compared to the results provided by the criterion of Mac Whirter. Then, the density of ionized species must be negligible compared to the density of neutral species. The minimum delay is 2 μs.

Finally, the time window must correspond to a period during which the plasma is optically thin. This condition has also been checked for the 2-10 μs window.

B. TRANSMISSION SPECTRUM OPERATION

The previous verifications have confirmed that the method is applicable (that the validity conditions of equation (S2) is met, in particular), the method is implemented and provides the following results.

Two emission lines are used for measuring stoichiometry: the carbon line at 247 nm, and the silicon line at 288.157 nm.

Indeed :

- the line at 247 nm of carbon is the only one recorded by the spectrometer used;

the signal-to-noise ratio of the silicon line at 288.157 nm is better, and the repeatability of this line is better than that of the other silicon lines.

The characteristics of these lines as well as the temperature of the plasma are presented in the table below:

Element λ (nm) Aij * 10 ⁷ Ei (eV) Ej (eV) 9i ZT (K)

If 288.16 18.9 5.082 0.780 3 16, 87 14800

C 247.85 3.4 7.684 2,684 3 11, 51 14800

The choice of lines (and therefore tabulated values) for calculating stoichiometry are presented in the table below.

According to equation (S), the ratio of the intensities of the lines corresponding to two transitions of the two elements Si and C is directly proportional to the ratio of the total densities of these two elements in a given ionization state, the neutral state in the occurrence, the other ionization states being negligible, as has been verified. Also, the stoichiometry can be calculated. Recorded emission spectra are used for an initial delay T ₀ of 4.2 μs and a camera integration time Δt of 2 μs.

The results are as follows:

The measured stochiometry (right column) corresponds, with a measurement error, to the theoretical ratio between carbon and silicon (1, 2, 4, 8). The dependence of the results obtained with respect to the various parameters of the equation (S) was finally evaluated. The electronic temperature T _e is the parameter that can cause the highest relative errors on the stoichiometry calculation.

The impact of the variations of the electronic temperature T _e is given in FIG. 10. This figure shows the variations of the ratio Nc / Nsi for different values of T _θ , namely 14800K, 14800 -1480K and 14800 + 1480K, and for four types of manufactured particles, SiC, SiC ₂ , SiC ₄ , SiCs. A 25% error in determining stoichiometry occurs for an error of 1500 K in the determination of the electronic temperature.

Claims

1, Method for analyzing composite particles having a diameter of less than 1 μm, making it possible to determine the stoichiometry thereof, and in which, at least once, the following steps b1) and b2) are carried out: b1) a laser firing (10) on a stream of an aerosol containing the particles to be analyzed, so as to form a plasma (12), and b2) acquires at least one emission spectrum (18) from the plasma to the using a light collection device (66) connected to an optical spectrometer (68) equipped with a detector (70), the acquisition taking place between an initial instant TD and a final instant TF, these two moments being counted from the moment of the laser shot; and then, c) from the emission spectrum (s) (18) obtained, the stoichiometry of the particles is determined; the method being characterized in that step c) comprises the following operation;

the proportion S or S _a / _b is calculated between two elements a, b, by the equation (Sl):

in it, the ISI factor ⁸ ' ¹ to _t / N ⁵ ' ¹ _tot is determined by the equation (S2):

equations in which; a, b are the two elements whose ratio S _a / _bA ij is the transition considered for the element a, between the levels I and j, N ^a τ _ot is the total number of atoms of the element a , sum of the atoms N ^aI τo _t in the neutral state I, that is to say the degree of ionization 0, and of the atoms N ^{a <1I} τo _t at the first ionized state II, that is to say at the degree of ionization +1, I ^{a / I} ι _j is the total intensity corresponding to a line of element a in the neutral state, corresponding to the transition between states i and j, is the wavelength of the radiation emitted during this transition , Z ^{3 / I} is the partition function of the element a in the neutral state, g, and gk are the statistical weights respectively of the level i of the element a and the level k of the element b,

A ^ _1J is the probability of transition between the levels i and j for the element a in the neutral state,

E ^a ' ^! _ι is the energy of the upper level i for the element a in the neutral state, k is the Boltzmann constant, T _e is the electronic temperature.

2. Analysis method according to claim 1, wherein in step c), the factor N ^a ' ^π _t o _t / N ^{a / I} _tot of the equation (Sl) is determined by the equation (S3 ):

where m _e is the mass of the electron, h is the Planck constant, and

3, Method of analysis according to claim 1, wherein from the time T ₀ , the ionized forms of the elements constituting the particles are substantially in number at least ten times lower than the number of the same elements in the neutral state, whereby in step c) the factors N ⁹⁴ WN ³ ' ¹ "and

early of the equation (Sl) are considered negligible, and so, the proportion S ₃ / _b between two elements a, b, according to (Sl), verifies;

Sa / b = N% _o1 / NV, = N ^a 'WN ^b ' ^! early

4, Method of analysis according to any one of claims 1 to 3, wherein step c) comprises the following step: the stoichiometry is determined by choosing as value for a relative proportion (S _a / _b ) between two elements (a, b), the value for which is minimal a difference between the observed spectrum, and the spectrum simulated by a plasma simulator .

5. Method of analysis according to any one of claims 1 to

4, in which the sequence of laser firing steps, or step b1) and acquisition of at least one spectrum, or step b2) is repeated at least a hundred times, and step c) is performed using the information accumulated over all emission spectra (18) thus acquired.

6. Method of analysis according to any one of claims 1 to

5, wherein the plasma (12) is optically thin between times T _D and T _F.

7. Method of analysis according to any one of claims 1 to

6, in which the plasma (12) is at the local thermodynamic equilibrium between the instants Tp and T _F.

8. Method of analysis according to any one of claims 1 to

7, wherein in step c), each of the spectra used to determine the relative proportions of the different elements constituting the particles, is acquired continuously from time T ₀ and up to time T _F.

9. The method of analysis according to any one of claims 1 to

8, wherein the aerosol comprises at least two types of composite particles.

10. Method of analysis according to any one of claims 1 to

9, in which the plasma generated during a step b1) contains at least 300 particles, and preferably at least 1000 particles, whose radiation is acquired in step b2) in order to allow the determination of the particle stoichiometry. in step c).

11. Method of analysis according to any one of claims 1 to

10, wherein in step b1), the aerosol in which the plasma is generated contains at least ⁷ particles per cubic centimeter, the particle size being between 10 nm and 100 nm.

12. Analysis method according to any one of claims 1 to 11, wherein before step b1), a step a) takes place, during which the concentration of particles in the aerosol is adjusted upwards or downwards.

13. Analysis method according to any one of claims 1 to 12, used to determine the stoichiometry of nanoparticles with a maximum diameter of less than 100 nm.

14. The method of analysis according to any one of claims 1 to

13, used to determine the stoichiometry of particles of silicon carbides, titanium, boron; titanium oxides; or iron oxide Fe ₂ O ₃ .

15. The method of analysis of any one of claims 1 to

14, wherein in step b1), in the aerosol the particles are suspended in a monoatomic inert gas such as argon, helium, neon.

16. A process for producing composite particles with a diameter of less than 1 μm per laser pyrolysis, characterized in that it comprises the on-line control of the stoichiometry of the particles produced, by means of the analysis method according to any one of the Claims 1 to 15.

17. Analysis system for determining the stoichiometry of composite particles with a diameter of less than 1 μm, the system comprising:

a chamber interposed on a channel (46), the channel being able to convey an aerosol containing the particles to be analyzed;

- A laser (60), adapted to perform a firing (10) on the aerosol as it passes through the chamber, to locally form a plasma (12);

a device (66) for collecting the light emitted by the plasma;

an optical spectrometer (68), equipped with a detector (70) to which the collected light is directed;

calculating means (76) able, from the emission spectrum or spectra obtained, to determine the stoichiometry of the particles; the system being characterized in that the calculating means (76) is provided for calculating the stoichiometry as follows;

the proportion S or S _a / _b between two elements a, b is calculated using the equation (Sl):

wherein Ie factor N ^ tot / N ⁵ ^'1 is early determined by the equation (S2): κ _l A, I ja.ï ia, I is j * b, the

_M bj ^~~ rb, I ₂ b, the _{ι Δ 7b} has, I ^ex P _κ τ

^N ttoot ^{l ι} kkll ^To "kkll ^Z " ^{• • '} 9 Mii ^{A k} uij

equations in which: a, b are the two elements whose ratio S _a / _{b is determined} , ij is the transition considered for the element a, between the levels i and j, N ^a τot is the total number of atoms of the element a, sum of the atoms

N ^aI τot in the neutral state I, that is to say at the degree of ionization 0, and atoms

N ^a ' ^π τot at the first ionized state II, that is to say at the degree of ionization +1,

is the total intensity corresponding to a neutral line of the element a, corresponding to the transition between the states i and j, λ ^{a / I} ij is the wavelength of the radiation emitted during this transition,

Z ^a ' ¹ is the partition function of the element a in the neutral state, g, and gk are the statistical weights respectively of the level i of the element a and the level k of the element b, is the probability of transition between the levels i and j for the element a, in the neutral state

E ^a 'j is the energy of the upper level i for the element a in the neutral state, k is the Boltzmann constant,

T ₆ is the electronic temperature.

18. An analysis system according to claim 17, wherein the calculating means determine the factor N ^a ' ^π _t o _t / N ³ ' ¹ _tot of the equation (Sl) using the equation (S3);

in which m _e is the mass of the electron, h is the Planck constant, and

Ej _0n is the ionization energy of an atom of the element a passing from the neutral state I to the state once ionized IL

19. The analysis system according to claim 17, wherein the light collection means are provided to collect the light only from the initial moment T _{D /} from which the ionized forms of the elements constituting the particles are substantially in number at least ten times smaller than the number of the same elements in the neutral state, whereby in step c) the factors

_tot of the equation (Sl) are considered negligible, and thus the proportion S _a / _b between two elements a, b, according to (Sl), satisfies:

early

20. Analysis system according to any one of claims 17 to

19, wherein the channel (46) is installed between two points (B, C) of at least one transport pipe (40,43) of the aerosol, a pressure difference between these two points inducing the movement of the aerosol in the canal.

21. Analysis system according to any one of claims 17 to

20, further comprising a gas supply channel, for adding gas into the aerosol upstream of the chamber to reduce the concentration of particles in the aerosol.