FR3072968A1 - USE OF A COLLOIDAL DISPERSION AS A GPF REGENERATION ADDITIVE - Google Patents

Abstract

The present invention relates to the use of a colloidal dispersion comprising particles as a regeneration additive of a particulate filter of a gasoline engine with direct or indirect injection. The invention also relates to a regeneration process of said particulate filter.

Description

Use of a colloidal dispersion as an additive for regenerating a GPF

Technical area

The present invention relates to the use of a colloidal dispersion as a regenerative additive for a particulate filter of a petrol engine with direct or indirect injection. The invention also relates to a method for regenerating said particle filter.

Technical problem

The automobile market is currently experiencing the commercial development of new so-called direct injection petrol engines which have reduced petrol consumption and better driving pleasure (low-end torque). To improve performance, these new engines operate with high fuel injection pressure and fuel injectors with the correct geometry. The air / petrol mixture is obtained in the combustion chamber. In fact, this particular functioning can lead to heterogeneities in the combustion and the formation of carbonaceous particles for which environmental standards, whether European, Asian or American, impose increasingly strict limits. This particular operation also generates the formation of deposits at the level of the gasoline injectors themselves and downstream from them. This fouling has an impact on combustion and therefore on primary emissions (that is to say the emissions at engine output), in particular those of carbonaceous particles, which become significant. A similar phenomenon with indirect injection gasoline combustion systems has also been observed.

To meet the regulations and the sustainability of emissions, we associate with gasoline engines with direct or indirect injection, a particulate filter adapted to these gasoline engines, also called GPF (Gasoline Particulate Filter). The operation of the GPF is known: the carbonaceous particles emitted by the engine accumulate on the filter and during the regeneration phase, (so-called purging), for example when the quantity of particles accumulated has reached a given threshold, the carbonaceous particles are burned inside the filter and released into the atmosphere as carbon dioxide. If the operation resembles that of a DPF, it differs from it, however, because unlike diesel engines, petrol engines operate with a lower air / fuel ratio, which therefore limits the presence of oxygen downstream of the combustion chamber. Consequently, it proves more difficult to oxidize the carbonaceous particles which would be deposited on the walls of the GPF.

It is therefore necessary to develop a solution which aims to effectively burn the carbonaceous particles which are deposited on a particle filter of a gasoline engine with direct or indirect injection and this whatever the driving configuration of the vehicle ( altitude, cold climate, ...).

Brief description of the invention

The invention relates to the use of a colloidal dispersion comprising particles consisting of an oxide and / or a hydroxide and / or an oxyhydroxide:

- iron and / or cerium; or

- of cerium and of an element E chosen from the group formed by the following elements Al, Cu, Ti, Zr, La, Pr, Nd and Y;

dispersed in a liquid organic phase, as a regenerative additive for a particulate filter of a gasoline engine with direct or indirect injection.

The term colloidal dispersion is used to describe particles dispersed in a liquid phase and having a size between 1 and 100 nm.

The invention also relates to a process for the regeneration of a GPF of an internal combustion engine operating on gasoline, consisting in using a gasoline to which a colloidal dispersion comprising particles consisting of an oxide and / has been added. or a hydroxide and / or an oxyhydroxide:

- iron and / or cerium; or

- of cerium and of an element E chosen from the group formed by the following elements Al, Cu, Ti, Zr, La, Pr, Nd and Y.

The invention therefore also relates to a process for regenerating a GPF of an internal combustion petrol engine consisting in using a petrol in which are dispersed particles consisting of an oxide and / or a hydroxide and / or a oxyhydroxide:

- iron and / or cerium; or

- of cerium and of an element E chosen from the group formed by the following elements Al, Cu, Ti, Zr, La, Pr, Nd and Y.

Regeneration consists in burning the carbonaceous particles which have deposited on the walls of the GPF. The invention can be applied to a direct injection petrol engine in which the petrol is injected directly into the combustion chamber of the engine or else to an indirect injection petrol engine in which the petrol is injected upstream in the intake manifold upstream of the intake valve. In engines with direct or indirect injection, fuel injection is frequently controlled by an electronic unit according to predefined parameters (e.g. engine speed, engine temperature, engine load, etc.).

Technical background

Dispersions based on iron or cerium are already known as additives for diesels under the term FBC for Fuel Borne Catalyst. The particles of iron or cerium which are present in the diesel have the function of catalyzing in the presence of oxygen the combustion of the soot which is deposited on a particle filter of the DPF type (Diesel Particulate Filter). Thus, for example, WO 2008/116550 describes the use of a colloidal dispersion of an iron compound used to reduce the fouling of injectors of a diesel engine. The use and the process described in the present application are therefore different because the particle filter is different and works differently from a DPF.

detailed description

The particles consist of an oxide and / or a hydroxide and / or an oxyhydroxide:

- iron and / or cerium; or

- of cerium and at least one element E chosen from the group formed by the following elements Al, Cu, Ti, Zr, La, Pr, Nd and Y.

It can be considered that the composition of the particles corresponds essentially to an oxide and / or a hydroxide and / or an oxyhydroxide of iron and / or of cerium or of cerium and of an element E. The term essentially means that the particles consist of 'an oxide and / or hydroxide and / or a hydroxide and can also contain residual compounds, resulting from the process of preparation of the particles. For example, in the case of precipitation of the particles from an aqueous nitrate or chloride solution, the particles may contain nitrate or chloride ions. Similarly, in the case where an iron or cerium complexing agent is used in the precipitation as taught in application WO 2003/053560, the particles may contain residues of said complexing agent.

In the case of mixed particles based on iron and cerium, the proportion of iron in the particles can vary from 0.5% to 50%, more particularly from 0.5% to 25%, this proportion being expressed by weight d iron oxide relative to the total weight of the particles. In the case of mixed particles based on E and on cerium, the proportion of the element E or the total proportion of the elements E in the particles can vary from 0.5% to 50%, more particularly from 0.5% to 25%, this proportion being expressed by weight of the oxide or oxides of E relative to the total weight of the particles. The proportion of cerium in the mixed particles can vary from 50% to 99.5%, more particularly from 75% to 99.5%, this proportion being expressed by weight of cerium oxide relative to the total weight of the particles. It will be noted that in the case where E is zirconium, hafnium is generally also present. These proportions are given by weight of oxide unless otherwise indicated. It is considered for these calculations that the cerium oxide is in the form of ceric oxide (CeO2) and that the other elements are in the following forms: AI2O3, Fe ₂ O3, CuO, T1O2, ZrO2, HfO2, La ₂ O3, Pr ₆ On, Nd ₂ O3, Y2O3. The proportions of the elements can be obtained using the usual analysis techniques in laboratories, in particular X-ray fluorescence.

In the dispersion, the particles are dispersed in a liquid organic phase. Preferably, the liquid organic phase is chosen to be compatible with the gasoline used. It should be noted, however, that the problem of compatibility or interference with petrol is generally not to be feared when the amount of dispersion which is added to petrol is small, which is generally the case. The liquid organic phase can be an apolar hydrocarbon, that is to say one which exhibits a resulting dipole moment zero. As an example of a liquid organic phase, there may be mentioned aliphatic hydrocarbons such as hexane, heptane, octane, nonane; cycloaliphatic hydrocarbons such as cyclohexane, cyclopentane or cycloheptane; aromatic hydrocarbons such as benzene, toluene, ethylbenzene, xylenes, liquid naphthenes. Also suitable are petroleum fractions containing a mixture of iso- and cyclo-paraffinic hydrocarbons, such as Isopar® or Isane®. For example, it is possible to use Isopar® L, Isane® 175 or Isane® 185. Also petroleum fractions containing a mixture of alkylbenzenes, in particular dimethylbenzene and tetramethylbenzene, are suitable. Solvesso®.

The particles of the dispersion according to the invention can be amorphous or crystallized.

According to another characteristic of the invention, at least 85%, more particularly at least 90% and even more particularly at least 95% of the particles of the dispersion are primary particles. By primary particle is meant a particle which is perfectly individualized and which is not aggregated with another or more other particles. This characteristic can be demonstrated by the MET counting method (high resolution transmission electron microscopy) described below. The% given here is a% by number.

The particles of the dispersion can have at least one of the following characteristics:

when the particles are crystallized, an average size d _D Rx determined by the X-ray diffraction technique less than or equal to 10 nm;

an average size d _M ET determined using the MET less than or equal to 10 nm;

a hydrodynamic diameter D _h measured by dynamic light scattering, less than or equal to 30 nm.

düRx corresponds to the size t of the coherent domain calculated according to the Scherrer model from the width of one or more of the most intense diffraction peaks of the oxide. It is possible to use several characteristic peaks (2 or

3) to calculate d _D Rx. In this case, d _D Rx corresponds to the arithmetic mean of the corresponding sizes t. According to Scherrer's model, a size t is determined by the following formula:

Λ · λ t: size at angle 2Θ (theta);

k: form factor equal to 0.89;

λ (lambda): wavelength of the incident beam 1.54 A;

H: width at mid-height of a characteristic peak;

s: width due to the defect of the instrumental optics which depends on the instrument used and the angle Θ (theta);

Θ: Bragg angle.

The instrumental width s can be determined in a manner known to those skilled in the art by a LaB ₆ analysis.

The particles may consist of an oxide, in particular an iron oxide, in particular a crystallized iron oxide. When the iron oxide is magnetite and / or maghemite and we can retain the diffraction peaks (440) described respectively in files 01-088-0315 and 00-039-1346 and of the ICDD (International Center for Diffraction Data). When the iron oxide is not very crystallized, the broad diffraction peak situated between 35 ° and 61 ° can be retained.

The particles may consist of a cerium oxide. In this case, we can retain the diffraction peaks described in file 01-089-8436 of the ICDD. We can in particular retain the peaks at the following angles: 28.5 ° ± 0.1; 47.5 ° ± 0.1 and 56.3 ° ± 0.1.

When the particles are based on iron and cerium or E and cerium, in particular when the particles consist of a mixed oxide of iron and cerium or of a mixed oxide of cerium and of the element (s) E, the characteristic peaks which can be retained may correspond to those of iron or cerium oxide or of element E or may be offset from them.

The DRX analysis can be carried out for example on a commercial device of the X’Pert PRO MPD PANalytical type, composed in particular of a Θ-Θ goniometer, allowing the characterization of liquid samples. The sample remains horizontal during the acquisition and the source and the detector move. This installation is controlled by the X’Pert Datacollector software supplied by the manufacturer and the exploitation of the diffraction diagrams obtained can be carried out using the X’Pert HighScore Plus software version 2.0 or higher (PANalytical supplier).

cImet is calculated from a distribution of particle diameters determined using the MET. The particles have a d _M ET less than or equal to 10 nm, d _M ET being calculated from a distribution of particle diameters determined using the TEM. The method for obtaining the distribution consists in measuring the diameter of at least 300 particles on one or more electron microscopy plate (s) (after having deposited the dispersion on a membrane and having allowed the liquid organic phase to evaporate). The enlargement of the microscope which is chosen must make it possible to clearly distinguish the images of the particles on a photograph. The magnification can for example be between 50,000 and 1,000,000. The diameter of a particle which is retained is that of the minimum circle making it possible to circumscribe the entire image of the particle as it is visible on a MET snapshot. The term minimum circle (in English, minimal enclosing circle) has the meaning given to it in mathematics and represents the circle of minimum diameter allowing to contain a set of points of a plane. Only the particles of which at least half of the perimeter is defined are retained. You can use ImageJ software to carry out the treatment more simply; This open source software was originally developed by the American institute NIH and is available at the following address: http://rsb.info.nih.gov or http://rsb.info. nih. gov / ij / download. HTMI. After determining the diameters of the particles retained by the above method, said diameters are grouped into several particle size classes ranging from 0 to 300 nm, the width of each class being 1 nm (it being understood that for dispersions having a narrow distribution , some of the classes may be empty). The number of particles in each class is the basic data to represent the number distribution (cumulative). From the distribution, the mean diameter d _M ET is determined which corresponds to the median diameter as conventionally understood in statistics. d _M ET is such that 50% of the particles (by number) taken into account on the MET plate (s) have a diameter smaller than this value.

The hydrodynamic diameter D _h (median diameter) is measured by dynamic light scattering from a number distribution of particle sizes. The measurement is made in the liquid organic phase of the dispersion.

As examples of colloidal dispersions, mention may be made of those described in the following patent applications and patents: US 6,210,451, EP 671205, WO 97/19022, WO 01/10545 and WO 03/053560, the latter two notably describing dispersions based on of cerium and iron compounds respectively and additionally containing an amphiphilic agent. Applications WO 2012/084838 and WO 2012/084851 also describe dispersions of an iron compound in crystallized form which can also be used. Patent EP 2 129 751 B1 describes iron oxide particles on which a long chain polycarboxylic acid is grafted.

The particles may consist of an oxide and / or an hydroxide and / or an iron oxyhydroxide. These particles can be prepared according to the teaching of international application WO 03/053560.

The particles may consist of an oxide and / or a hydroxide and / or a cerium oxyhydroxide. These particles can be prepared according to the teaching of US 6,210,451 or EP 0671205. More particularly, the particles can be based on cerium dioxide and be in the form of crystallite agglomerates of which the d80 (number distribution) is at most equal to 10 nm, d80 being determined by counting by high resolution transmission electron microscopy.

The particles can be in the form of cerium oxide. These particles can be prepared according to the teaching of international applications WO 97/19022 or WO 01/10545.

To ensure stability of the dispersion, it advantageously comprises at least one amphiphilic agent whose function is to stabilize the particles. The amphiphilic agent can be a carboxylic acid comprising from 7 to 50 carbon atoms, preferably from 7 to 25 carbon atoms. The carboxylic acid can be a monoacid or a diacid. This acid can be linear or branched. It can be chosen from aryl, aliphatic or arylaliphatic acids, optionally carrying other functions provided that these functions are stable in the media where it is desired to use the dispersions according to the present invention. Thus, it is possible to use, for example, aliphatic carboxylic acids, whether natural or synthetic.

Examples of acids that may be mentioned include tall oil fatty acids, soybean oil, tallow, linseed oil, oleic acid, linoleic acid, stearic acid and its isomers. , in particular isostearic acid, pelargonic acid, capric acid, lauric acid, myristic acid, dodecylbenzenesulfonic acid, 2-ethylhexanoic acid, naphthenic acid, acid hexoic. Oleic or isostearic acid are two acids which may be suitable for stabilizing particles. It may also be succinic acids substituted by polybutenyl groups, for example the PIBSA of formula (I) in which PIB denotes a polybutenyl group:

O

GDP. AT.

OH L ^ OH ⁰ (I)

The amphiphilic agent can also be the compound of formula (II):

C "3 (II) or the diacid derived from the compound of formula (II).

In the case of amphiphilic agents of formula (I) or (II), the agent can be chemically grafted onto the particles. For this, it is possible to react in an organic solvent at reflux a dispersion of an iron compound which is prepared from iron (III) chloride and sodium bicarbonate and the compound of formula (I), of formula ( II) or the diacid derived from said compound of formula (II), following the teaching of Example 1 of EP 2129751 B1.

Mixtures of acids can also be used, in particular mixtures of acids which contain chain length distributions. For example, Prisorine 3501 from Croda corresponds to a mixture of C16-C22 acids.

The amphiphilic agent can also be chosen from polyoxyethylenated alkyl ether phosphates. Here we mean the phosphates of formula:

Ri-O- (CH ₂ -CH ₂ -O) nP (= O) (OM) ₂ or alternatively the polyoxyethylenated dialkyl phosphates of formula: R ₃ -O- (CH ₂ CH ₂ O) _n -P (X) (= O) (OM) with X = R ₂ -O- (CH ₂ CH ₂ O) _n formulas in which:

Ri, R ₂ , R _{3i, which are} identical or different, represent a linear or branched alkyl group, which can comprise from 2 to 20 carbon atoms; a phenyl radical; an alkylaryl radical, more particularly an alkylphenyl radical, in particular having an alkyl chain group of 8 to 12 carbon atoms; an arylalkyl group, more particularly a phenylaryl radical;

n is an integer representing the number of ethylene oxide between 0 and 12;

M represents a hydrogen, sodium or potassium atom.

The radical Ri may in particular be a hexyl, octyl, decyl, dodecyl, oleyl, nonylphenyl radical.

As an example of this type of amphiphilic compound, mention may be made of those sold under the brands Lubrophos® and Rhodafac® sold by Rhodia and in particular the products below:

- poly-oxy-ethylene alkyl (Cs-C-10) ethers phosphates Rhodafac® RA 600;

- poly-oxyethylene tridecyl ether phosphate Rhodafac® RS 710 or RS 410;

- poly-oxy-ethylene oleocetyl ether phosphate Rhodafac® PA 35;

- poly-oxy-ethylene nonylphenyl ether phosphate Rhodafac® PA 17;

- poly (oxy-ethylene nonyl (branched) ether phosphate Rhodafac® RE 610.

The amphiphilic agent can also be chosen from polyoxyethylenated alkyl ether carboxylates of formula: R ₄ - (OC ₂ H ₄ ) _n -O-Rs in which R ₄ is a linear or branched alkyl group which can especially comprise 4 to 20 atoms of carbon, n is an integer which can range for example up to 12 and R ₅ is a residue of carboxylic acid such as for example -CH ₂ COOH. By way of example, mention may be made for this type of amphiphilic compound of those sold under the brand AKIPO® by Kao Chemicals.

In addition to the particles described above, the dispersion can also comprise at least one other additive chosen from the group formed by corrosion inhibitors, additives improving the lubricating power, detergents, defoamers, antifreezes, antioxidants, dyes, stabilizing additives, additives improving the octane number. These latter additives are used to prevent wear or seizure of high pressure pumps in particular and injectors. This other additive can be added directly to the colloidal dispersion. According to a variant, the combination of the particles and the other additive can be obtained by separately introducing said additive into the gasoline.

Thus, the dispersion includes:

the particles as previously described, in particular the particles of iron oxide, of cerium oxide, of mixed oxide of iron and of cerium or of mixed oxide of cerium and of element E;

- at least one amphiphilic agent;

- the liquid organic phase;

- optionally an additive chosen from the group formed by corrosion inhibitors, additives improving the lubricating power, detergents, defoamers, antifreezes, antioxidants, dyes, stabilizing additives, additives improving the index of octane.

According to one embodiment, the dispersion does not include the additive. According to another embodiment, the dispersion consists of the particles, at least one amphiphilic agent and a liquid organic phase.

The dispersion is used as a regenerative additive for a particulate filter of a gasoline engine with direct or indirect injection. This particulate filter (commonly called GPF) is placed on the exhaust line and has the function of reducing the number of carbonaceous particles emitted by the engine of the vehicle, this in order to meet environmental standards on the release of carbon particles.

The GPF is a porous device which makes it possible to reduce, by a filtration mechanism, the emissions of carbonaceous particles below a regulatory limit, for example 6.0 x 10 ⁺¹¹ particles / km when considering the vehicle or 6.0 x 10 ⁺¹¹ particles / kW.h when only the engine is considered, especially in dynamic cycle (eg according to the World Harmonized Transient Cycle - WHTC - or the recent European Transient Cycle - ETC). This particulate filter is therefore suitable for the filtration of carbonaceous particles from the combustion of petrol and it differs from the conventional particle filters used to filter soot particles from the combustion of diesel (also commonly called DPF). The GPF can be made of a porous ceramic, a ceramic or metallic foam, or a tangle of ceramic or metallic wires. The geometry of the filter (size, filtration surface, porosity and pore size distribution) is variable and depends on the solution chosen by the manufacturer, the constraints of which are as follows:

- a reduction in emissions of carbonaceous particles below the regulatory value;

- a limitation of the fuel consumption of the vehicle;

- maintain filter efficiency even in the presence of residues from the consumption of oil or other fluids and other components used by the engine.

The particles of the dispersion have the function of catalyzing the combustion of the carbonaceous particles which have deposited on the walls of the filter. The particles of the dispersion make it possible to burn the carbonaceous particles at a lower temperature and / or for a shorter regeneration time than in the absence of particles of the dispersion. Regeneration is implemented in order to burn all the carbonaceous particles present on the walls of the filter. Regeneration is generally caused, i.e. triggered by an engine injection computer when a parameter representative of the quantity of carbonaceous particles in the GPF is reached. Each car manufacturer can choose the appropriate trigger mode. Thus, to trigger regeneration, gasoline can be injected at the GPF inlet or in the engine cylinders at the end of the compression time so that the gasoline does not participate in combustion and is evacuated at the exhaust. It is also possible to adjust the operation of the engine by adjusting the ignition and / or the phasing of the injection of petrol into the combustion cycle so as to degrade the combustion efficiency and thus increase the heat losses on exhaust. The duration of the regeneration is therefore variable, it being understood that it is preferable that this duration be as short as possible. It is also possible that the GPF works continuously and that the carbonaceous particles are burned as they accumulate on the walls of the filter.

In general, the temperature of the exhaust gases for a petrol engine is higher than for a diesel engine. The exhaust gas temperatures of a petrol engine can be in the range of 550 ° C to 670 ° C. During regeneration, the temperature of the GPF must be sufficient for the carbonaceous particles to be effectively burned, whether for triggered regeneration or for continuous regeneration. Preferably, for efficient combustion of carbonaceous particles, the temperature of the GPF is at least 200 ° C., or even at least 400 ° C., or even at least 450 ° C. The GPF can be placed in several places of the exhaust line. It can be in close-coupled position (that is to say close to the engine) or sub-floor (that is to say in a position on the exhaust line more distant than the close-coupled position) . The closecoupled position is advantageous because thus, the GPF can benefit more efficiently from the caloric energy provided by the combustion gases.

It will be noted that, in addition to the GPF, the exhaust line can also include at least one other device making it possible to reduce other pollutants such as CO, unburnt hydrocarbons or NO _X. This other device can be placed on the exhaust line before or after the GPF in the direction of circulation of the exhaust gases. Fig. 1 shows several examples of combinations of such a TWC (Three-Way Catalyst or 3-way catalyst) type device with the GPF.

Methods of introducing dispersion

Addition of dispersion in petrol: dispersion can be added to petrol in several ways. It can be added to petrol continuously or discontinuously. Dispersion is added to the fuel either at the fuel tank or in the low pressure fuel line (e.g. in the outward line, the return line, or at the fuel filter). The dispersion is contained in an on-board tank. The discontinuous introduction corresponds to an intermittent addition of the dispersion in gasoline, which saves the dispersion. It can be added to petrol using a metering pump which can be controlled by an on-board electronic system (ECU in English). The addition of the dispersion is controlled according to several parameters of the ECU such as for example the quantity of carbonaceous particles in the GPF, the temperature of the GPF, the moment when the regeneration is triggered, ...

Dispersion can also be added manually and batchwise to the fuel tank. The dispersion is then contained in doses and pods which are generally sold in networks accessible to the general public.

Whatever the method of adding dispersion to petrol, the proportion of dispersion in petrol is generally less than or equal to 100 ppm, or even less than or equal to 50 ppm, this proportion being expressed by weight by weight particles constituting the dispersion with respect to petrol just before its injection into the combustion chamber. This proportion can more particularly be between 1 and 50 ppm.

Introduction of the dispersion into the exhaust line: the dispersion can also be introduced into the exhaust line, that is to say between the exhaust manifold and the GPF, from an on-board tank containing the dispersion. The point of introduction of the dispersion can be optimized to reduce the fouling of mechanical parts present on the exhaust line, such as EGR valves, or carbon deposits which could poison the TWC catalysts located upstream of the GPF. . The dispersion can be introduced by an injection means, such as for example an injector. This method of introduction has the advantage of effectively using the dispersion since the dispersion can be introduced only when necessary, which makes it possible to limit consumption. Again, the dispersion can be introduced into the exhaust line continuously or discontinuously. Again, it can be added to petrol using a metering pump which can be controlled by an ECU. The introduction is controlled according to several parameters of the ECU such as for example the quantity of carbonaceous particles in the GPF, the temperature of the GPF, the moment when the regeneration is triggered, ...

The petrol engine runs on a lower oxygen content than a diesel engine. Generally, the stoichiometric air / fuel mass ratio is 14.7: 1. In operation, the ratio can be between 10: 1 and 70: 1. The factor λ can be between 0.9 and 1.10.

The engine fitted with a GPF operates with pressurized fuel injection. The fuel injection pressure can be at least 100 bar. This pressure can reach 500 bar, a current technical limitation but which could become higher with future technological progress.

Examples

We start by impregnating carbonaceous particles (soot) with a colloidal dispersion of CeO2 in water (15% by weight of Ce oxide relative to the soot). The combustion temperature of the soot is then determined using thermogravimetric analysis (ATD-ATG). For this, a TGA / DCS3 + LF1100 thermobalance from Mettler is used, connected directly at the outlet of the furnace to an IR spectrometer from Brüker for analyzing the combustion gases.

The program used for the ATG is as follows: temperature rise from ambient to 150 ° C at 5 ° C / min; isothermal at 150 ° C for 1 h; temperature rise from 150 ° C to 1000 ° C at 10 ° C / min. The sample is swept by a gas at a flow rate of 50 mL / min which is either nitrogen or a nitrogen mixture + 2% O2 (representative of the composition of an exhaust gas from a petrol engine which does not contain than a few hundred ppm O2). The sample is contained in a 70 μl alumina crucible with cover. We use mathematical compensation for Archimedes' thrust (equivalent to subtracting a blank test under the same conditions).

Fig. 2 represents the weight loss in% relative to the temperature in ° C for two samples under nitrogen sweep:

- soot without cerium oxide (white);

- the same soot with cerium oxide.

Fig. 3 represents the weight loss in% relative to the temperature in ° C for two samples under scanning of the nitrogen + O2 mixture:

- soot without cerium oxide (white);

- the same soot with cerium oxide.

These two figures show that the combustion of soot is initiated at a lower temperature in an atmosphere poor in O2.

Claims (13)

1. Use of a colloidal dispersion comprising particles consisting of an oxide and / or a hydroxide and / or an oxyhydroxide: - iron and / or cerium; or - of cerium and of an element E chosen from the group formed by the following elements Al, Cu, Ti, Zr, La, Pr, NdetY; dispersed in a liquid organic phase, as a regenerative additive for a particulate filter of a gasoline engine with direct or indirect injection. 2. Use according to claim 1 in which the particles have at least one of the following characteristics: an average size d _D Rx determined by the X-ray diffraction technique less than or equal to 10 nm; an average size d _M ET determined using the MET less than or equal to 10 nm; a hydrodynamic diameter D _h measured by dynamic light scattering, less than or equal to 30 nm. 3. Use according to one of the preceding claims in which the dispersion comprises: - the particles ; - at least one amphiphilic agent; - a liquid organic phase; - optionally an additive chosen from the group formed by corrosion inhibitors, additives improving the lubricating power, detergents, defoamers, antifreezes, antioxidants, dyes, stabilizing additives, additives improving the index of octane. 4. Use according to claim 3 wherein the amphiphilic agent is a carboxylic acid comprising from 7 to 50 carbon atoms, preferably from 7 to 25 carbon atoms. 5. Use according to claim 3 or 4 wherein the amphiphilic agent is chosen from tall oil fatty acids, soybean oil, tallow, linseed oil, oleic acid, linoleic acid, l stearic acid and its isomers, in particular isostearic acid, pelargonic acid, capric acid, lauric acid, myristic acid, dodecylbenzenesulfonic acid, 2-ethylhexanoic acid, naphthenic acid, hexoic acid. 6. Use according to one of the preceding claims in which the particles consist of iron oxide, cerium oxide, a mixed oxide of iron and cerium or a mixed oxide of cerium and of the item (s) E. 7. Use according to one of the preceding claims, characterized in that at least 85%, more particularly at least 90% and even more particularly at least 95% of the particles of the dispersion are primary particles. 8. Use according to one of the preceding claims in which the particles are amorphous or crystallized. 9. Use according to one of the preceding claims in which: - the dispersion is added to the petrol continuously or discontinuously; or - the dispersion is introduced into the exhaust line. 10. A method of regenerating a GPF of an internal combustion engine running on petrol consisting in using a petrol to which a colloidal dispersion has been added as described in one of claims 1 to 8. 11. A method of regenerating a GPF of an internal combustion engine using gasoline, comprising using a gasoline in which particles are dispersed as described in one of claims 1 to 8. 12. Method according to claim 10 or 11 in which: - the dispersion is added to the petrol continuously or discontinuously; or - the dispersion is introduced into the exhaust line. 13. The method of claim 10 to 12 wherein the GPF temperature is at least 200 ° C, or even at least 400 ° C, or even at least 450 ° C.