EP4162520A1

EP4162520A1 - Facility and method for distributing a gas mixture for doping silicon wafers

Info

Publication number: EP4162520A1
Application number: EP21728917.2A
Authority: EP
Inventors: Vanina TODOROVA; Hervé Dulphy
Original assignee: Air Liquide Electronics Systems SA
Current assignee: Air Liquide Electronics Systems SA
Priority date: 2020-06-05
Filing date: 2021-05-27
Publication date: 2023-04-12
Also published as: WO2021244946A1; KR20230020436A; FR3111086A1; TW202202653A; US20230285911A1; CN115699285A; JP2023533432A

Abstract

Disclosed is a facility for distributing a gas mixture to a silicon wafer doping unit, comprising a doping gas source (1), a carrier gas source (2), a mixing device (3) connected to the doping gas container (1) and the carrier gas source (2), a first flow rate regulating member (41) and a second flow rate regulating member (42) for regulating the flow rates of the doping gas (1) and carrier gas (2) to the mixing device (3), a control unit (5) for controlling the first and second flow rate regulating members (41, 42) so as to adjust the first flow rate setpoint (D1) and the second flow rate setpoint (D2) in proportions determined according to at least one target content (C1, C2) of the mixture of the doping gas (1) and/or the carrier gas (2), a buffer tank (7), a distribution line (6) for distributing the mixture to a doping unit (10) with a consumption rate (DC), at least one measurement sensor (8) for measuring a physical quantity, the variation of which is representative of a variation in the consumption rate (DC), and for supplying a first measurement signal, the control unit (5) being connected to the sensor (8) and configured to produce a first control signal on the basis of the first measurement signal, the flow rate regulating members (41, 42) being configured to adjust the first and second flow rate setpoints (D1, D2) in response to the first control signal.

Description

Installation and method of distributing a gas mixture for doping silicon wafers

The present invention relates to an installation for distributing a gas mixture intended for use by a unit for doping silicon wafers. The installation allows distribution of the mixture directly to the site of use as well as an adjustment of the mixture flow rate produced by the installation according to the flow rate consumed by the consumer unit. The invention also relates to an assembly for doping silicon wafers comprising such an installation.

In particular, an installation and a method according to the invention are intended to distribute mixtures of pure gases or of premixes of gases, in particular to distribute mixtures of so-called carrier gases and of so-called doping gases.

It should be noted that the terms "doping unit" can mean both a single doping entity and several entities supplied in parallel by the gas mixture, in particular several entities arranged downstream of a junction box.

The present invention applies in particular to the doping of silicon wafers in the process of making semiconductors.

In the process of manufacturing integrated circuits for electronics, semiconductor manufacturing technologies are mainly based on the intrinsic modification of the matrix comprising the silicon atoms by inserting so-called doping elements therein, in order to make the silicon semi -driver. Known doping elements are, for example, germanium, phosphorus, arsenic, antimony, boron, gallium, aluminum.

In the most widely used doping processes, with phosphorus or boron for example, the silicon wafers are introduced into a furnace and brought to a temperature generally between 800 ° C and 1200 ° C. Mixtures of doping gases and carrier gases are brought into the chamber of the furnace. The role of the carrier gas is to transport the doping gas on the surface of the silicon wafer.

Usually, gas mixtures are packaged in compressed or liquefied form in gas cylinders. The filling of a gas cylinder is carried out in sequential mode, the constituents of the mixture being introduced one after the other into the bottle. For each constituent, the quantity of gas introduced into the bottle is checked, either by monitoring the pressure in the bottle during and after the introduction of the constituent, or by weighing the bottle during the introduction of the constituent. Such an installation for conditioning gas mixtures is described in particular in document WO2010 / 031940A1.

In order to guarantee the user the reliability and reproducibility of the performance and / or the results provided by the gas consuming unit, it is necessary to produce gas mixtures with high precision on the concentrations of each component. Depending on the applications, the maximum tolerance for variation of the effective values of the concentrations with respect to the target values may be less than 1% (% relative), or even less than 0.5% or even less than 0.1%. Such tolerances are all the more difficult to comply with as the number of constituents is large and / or their contents are low.

This is the case in particular with doping mixtures for the manufacture of integrated circuits which involve relatively low mixing flow rates and dopant gas contents. This requires further improving the control of doping gas contents and ensuring their accuracy and stability, a control which turns out to be even more critical due to the nature of the doping gases, the constituents of which are potentially flammable, pyrophoric and / or. or toxic.

Depending on the precision required, current conditioning methods may prove insufficient. In particular, the manometric conditioning by pressure control offers a precision which is intrinsically limited by the precision of the pressure sensor and by the variations in temperature which influence the calculation of the quantity of gas. Added to the uncertainty in the concentration values of the gas mixture produced are the differences in concentrations between the mixtures packaged in different bottles. Such deviations can significantly vary the results produced by the consuming unit with each bottle change.

Gravimetric conditioning by weighing the constituents offers greater precision on the composition of the mixture but still requires a step-by-step process with filling of bottles. However, the use of bottles leads to a limited autonomy for the user with a stopping of the distribution which is difficult to predict when the consumption of the gas mixture varies. As the supply times for gas mixtures can be relatively long, the user must manage his stock of cylinders in order to ensure continuity of production.

In addition, the blends are bottled in packaging centers specially designed for this type of operation. The bottles must then be transported to their site of use, which requires dedicated logistics. Constraints linked to the transport of dangerous goods also arise when it comes to transporting gas mixtures with flammable, pyrophoric, toxic and / or anoxicants.

In addition, the operations of connecting / disconnecting the cylinders are tedious for the users and increase the risk of contaminating the gas mixture with ambient air. The bottles also require specific preparation before filling including cleaning, passivation, ...

The object of the invention is to alleviate all or part of the drawbacks mentioned above, in particular by proposing an installation for distributing a gas mixture intended to be used by a unit for doping silicon wafers, said installation making it possible to control the composition of the mixture, while offering continuity and flexibility of distribution, depending in particular on the needs at the point of consumption of the mixture.

To this end, the solution of the invention is an installation Installation for the distribution of a gas mixture suitable for and intended for use in a unit for doping silicon wafers, said installation comprising:

- a source of a doping gas,

- a source of a carrier gas,

a mixing device fluidly connected to the doping gas container and to the source of carrier gas, said mixing device being configured to produce at an outlet a gas mixture comprising the dopant gas and the carrier gas,

- a first flow regulator member and a second flow regulator member configured to respectively regulate the flow rate of the doping gas and the flow rate carrier gas flowing towards the mixing device according to a first flow rate instruction and a second flow rate instruction defining, in operation, a production flow rate of the gas mixture at the outlet of the mixing device,

- a control unit configured to control the first and second flow regulator members so as to adjust the first flow setpoint and the second flow setpoint according to respective proportions with respect to the production flow, said respective proportions being determined as a function of 'at least a target content of the gas mixture of the doping gas and / or the carrier gas,

- a buffer tank connected by an outlet pipe to the outlet of the mixing device on the one hand and to a distribution line on the other hand, the distribution line being configured to distribute the gas mixture to a platelet doping unit silicon with a consumption rate representative of a variable consumption of the gas mixture,

- at least one measurement sensor configured to measure a physical quantity the variation of which is representative of a variation in the consumption flow rate distributed by the distribution line and to provide a first measurement signal of said physical quantity, the control unit being connected to the measurement sensor and configured to generate a first control signal from the first measurement signal, the flow regulating members being configured to adjust the first flow setpoint and the second flow setpoint in response to said first control signal .

Depending on the case, the invention may include one or more of the characteristics set out below.

The installation comprises a first analysis unit arranged downstream of the buffer tank (and configured to analyze at least a respective content of the doping gas and / or the carrier gas of the gas mixture distributed by the supply line.

The installation comprises a first sampling duct connecting the first analysis unit to the supply line at a first sampling point and a first return duct connecting the first analysis unit to the supply line at a first restitution point, the restitution point being located downstream of the first point sampling on the supply line, a pressure reducing valve being mounted on the supply line between the first sampling point and the first return point, preferably the pressure reducing valve is mounted upstream of the measurement sensor.

The installation comprises a second analysis unit configured to measure at least one content of the doping gas and / or the carrier gas of the gas mixture produced at the first output of the mixing device and to supply consequently at least one second signal of measurement, the control unit being connected to the second analysis unit and configured to generate a second control signal from the second measurement signal and to modify the proportion of the first flow setpoint and / or the proportion of the second flow setpoint relative to the production flow rate in response to said second control signal.

The installation comprises a second sampling duct connecting the second analysis unit to the outlet pipe at a second sampling point and a second return pipe connecting the second analysis unit to the outlet pipe at a second outlet point. restitution, the restitution point being located downstream of the first sampling point on the outlet pipe, a regulator being mounted on the outlet pipe, between the second sampling point and the second restitution point.

The installation is configured to dispense a mixture having a dopant gas content of between 0.0001 and 50%, preferably between 0.05 and 30% (% by volume).

The dopant gas source contains germanium tetrahydride (GehU), phosphine (PH3), arsine (Ashh) and / or diborane (B2H6) and / or the carrier gas source contains hydrogen ( H2), nitrogen (N2) and / or argon (Ar).

The dopant gas source contains a gas premix formed of dopant gas and carrier gas.

The installation comprises a first loop for slaving the first and second flow setpoints on the first measurement signal supplied by the measurement sensor, said first loop comprising: - a first comparator arranged within the control unit and configured to generate at least a first error signal from the first measurement signal,

- a first corrector arranged within the control unit, in particular of the proportional, integral and derivative type, and configured to generate the first control signal from the first error signal,

- actuators of the first and second flow regulator members connected to the first corrector configured to receive the first control signal and move the first and second flow regulator members in respective positions in which the first flow setpoint and the second flow setpoint conform to the first control signal.

The installation comprises a second control loop of the respective proportions of the first flow setpoint and / or of the second flow setpoint relative to the production flow rate on the second measurement signal supplied by the second analysis unit, the second loop including:

a second comparator arranged within the control unit and configured to generate at least a second error signal from a comparison of the second measurement signal with at least one parameter chosen from: a target gas content dopant, a target content of the carrier gas,

- a second corrector arranged within the control unit, in particular of the proportional, integral and derivative type, and configured to generate the second control signal from the second error signal,

- the actuators of the first and / or second flow regulator members connected to the second corrector and configured to move the first and / or second flow regulator members in respective positions in which the proportions of the first flow setpoint and / or second setpoint flow rate versus production flow rate conforms to the second control signal.

The measuring sensor includes a flow sensor or flow meter configured to measure the consumption flow rate.

The first comparator is configured to generate at least a first error signal representative of a variation in the consumption rate and the first corrector is configured to generate a first control signal controlling a displacement of the first and second flow regulating members so that the first and second flow setpoints vary in the same direction as that of the flow rate variation.

The measurement sensor includes a pressure sensor configured to measure the pressure prevailing in the buffer tank.

The first comparator is configured to generate a first error signal representative of a variation in the pressure in the buffer tank and the first corrector is configured to generate at least a first control signal controlling a movement of the first and second regulator members. flow rate so that the first and second flow rate setpoints vary in a direction opposite to that of the pressure variation.

In addition, the invention relates to an assembly comprising a unit for doping silicon wafers comprising a furnace provided with an enclosure associated with heating means and a support arranged in said enclosure on which the wafers are installed, the furnace comprising means for introducing a mixture of doping gas and carrier gas into the enclosure, characterized in that it further comprises an installation according to the invention, said introduction means being fluidly connected to the line d supply of said installation.

The invention will now be better understood from the following detailed description given by way of illustration and not by way of limitation with reference to the appended figures described below.

Fig. 1 shows schematically the operation of an installation according to one embodiment of the invention.

Fig. 2 schematically shows a first servo loop according to one embodiment of the invention.

Fig. 3 shows an example of the change over time of the pressure in the buffer tank and the production flow rate of the installation.

Fig. 4 represents an example of controlled evolution of the content of a constituent of the gas mixture distributed by an installation according to one embodiment of the invention. Fig. 5 shows an example of the change over time of the gas mixture flow rate distributed by an installation according to one embodiment of the invention with the content of a constituent of the mixture measured during this change.

FIG. 1 represents an installation according to the invention comprising a source of dopant gas 1 and a source of carrier gas 2. These gases can be pure, simple or compound substances, or premixtures of several pure substances, in particular one. pure body diluted with another.

The term “dopant” is understood to mean a gas suitable and suitable for doping silicon in the field of semiconductors, that is to say a gas making it possible to introduce atoms of another material into the silicon matrix in order to modify the conductivity properties of silicon. As doping gas, it is possible in particular to use germanium tetrahydride (GeH4), phosphine (PH3), diborane (B2H6), arsine (AsH3).

The term “carrier” means a gas capable and suitable for transporting the doping gas to the silicon matrix, preferably a gas formed from one or more pure inert substances such as hydrogen (H2), nitrogen (N2). ) or argon (Ar).

It should be noted that the terms “doping gas” can cover a pure doping substance, a mixture of several pure doping substances or a premix comprising a pure doping substance diluted in a pure non-doping substance. Advantageously, the doping gas is formed from a pure doping substance diluted in another pure substance which is of the same nature as that forming the carrier gas. Since dopants are very reactive, some are usually stored at very low temperature, typically -30 ° C, in the liquid state, in order to ensure stability. By using a dopant gas premix diluted in carrier gas, the dopant is stored as a gas mixture, which ensures the stability of the dopant as well as better homogeneity.

Thus, it is possible to use a doping gas composed of a pure doping substance, in particular from 1 to 30% of pure doping substance, preferably from 1 to 15%, and of carrier gas for the remainder, to finally provide mixtures. dopants having dopant gas contents ranging from 0.0001% to 30% in the carrier gas. For example, the doping gas could comprise as pure doping substance B2H6 with a content of 10% in H2, then mixed with H2, to provide doping mixtures with contents of B2H6 in H2 ranging from 0.05% to 5%. Preferably, each of the gas sources is a receptacle containing said gas, in particular a gas cylinder, typically a cylinder capable of having a water volume of up to 50 L, or a set of cylinders connected together to form a frame. bottles or a tank with a larger capacity, in particular a capacity of up to 1000 L, such as a cryogenic storage tank or a tank arranged on a truck-trailer.

In particular, the source of doping gas is a container containing a doping gas and the source of carrier gas is a container containing a carrier gas.

Preferably, the sources dispense fluids in the gaseous state. Before distribution, fluids can be stored in the gaseous state, in the liquid state, i. e. liquefied gas, or two-phase liquid / gas. Preferably, in the case of a doping premix, it will be stored in the gaseous state.

Figure 1 illustrates the case where the installation is configured to produce a binary gas mixture, i. e. two-component, from two gas containers. Of course, an installation according to the invention could include more than two gas sources and produce mixtures with more than two constituents, in particular mixtures of ternary or quaternary gases.

Each of the doping gas 1 and carrier gas 2 receptacles is connected by a first pipe 21 and a second pipe 22 to respective first and second flow regulating members 41, 42. These are provided to regulate the doping gas flows. and carrier gas flowing towards the gas mixing device 3. Preferably, the pipes 21, 22 meet at a connection point 31 located upstream of the mixing device 3 to form a common portion of pipe connected to an inlet 32 of the mixing device. A mixture of dopant gas and carrier gas thus enters the device 3 in order to be further mixed and homogenized therein. Note that it is also possible for the pipes 21, 22 to open into two separate inlets 32a, 32b of the mixing device 3.

Preferably, each of the pipes 21, 22 is provided with a pressure reducing valve and a pressure sensor in order to measure and control the pressure prevailing in these pipes. The pressures of the doping gas and carrier gas can each be kept constant, typically at a value between 1 and 10 bar. Each flow rate regulator member 41, 42 can be any means configured to regulate, regulate, adjust the flow rate of a fluid to bring it to a flow rate value closest to the desired value.

Typically, the flow rate regulating members 41, 42 each comprise a flow rate sensor, or flow meter, associated with an expansion member, such as a valve, for example a valve with proportional adjustment. The valve can be pneumatic or piezoelectric, analog or digital. The valve comprises a movable part, typically at least one shutter, which is placed in the fluid flow and whose movement makes it possible to vary the passage section, and thus to vary the flow to bring it to the set value. In particular, the flow regulator members 41, 42 can be mass flow regulators comprising a mass flow sensor and a proportional control valve. Note that even if the regulation is based on a measurement of the mass of fluid, the set and measured flow rate values are not necessarily expressed in mass. Thus, a volume flow setpoint can be expressed as a percentage of opening of the proportional control valve, to which corresponds a voltage value to be applied to the control valve of the regulator. The conversion between percentage opening in mass or volume flow value is done by knowing the nominal value of the regulated flow for 100% opening.

According to an advantageous embodiment, the valve is piezoelectric. This type of valve offers high precision, good reproducibility, allowing the voltage applied to the valve to be monitored. Such valves are also insensitive to magnetic fields and radiofrequency noise. Their energy consumption is low with minimal heat generation. The metal-to-metal control surface reduces or even eliminates reactions with the gas. Finally, due to a relatively small flow control cavity volume, especially compared to that of a solenoid valve, it is possible to have rapid gas replacement and excellent dynamic response.

In practice, the first and second flow rate regulating members 41, 42 make it possible to regulate respectively the flow rate of the doping gas and the flow rate of the carrier gas entering the mixer 3 according to a first flow rate setpoint D1 and a second flow rate setpoint D2. At the outlet 33 of the mixing device 3, the gas mixture in an outlet pipe 23 with a production flow rate DP which corresponds, in the case of an installation with two gas sources, to the sum of the two flow rates D1 and D2 of dopant gas and carrier gas. If the installation comprises for example a source of a third gas, the flow rate DP will be the sum of the flow rates D1, D2, D3 regulated by the corresponding flow rate regulating members 41, 42, 43 towards the mixing device 3.

The installation according to the invention further comprises a control unit 5 which is connected to the first and second flow regulator members 41, 42 so as to control their operation, in particular so as to adjust the setpoint values D1, D2 for bring them to values which are determined and adapted according to the operating conditions of the installation.

To do this, the flow regulating members 41, 42 each advantageously comprise a closed loop system which is given flow setpoints by the control unit 5. These setpoints are then compared by the closed loop system with the values. measured by the flow rate regulating members 41, 42 and their positions are adjusted by said system accordingly to send the flow rates as close as possible to D1, D2 to the mixing device 3.

Advantageously, the control unit 5 comprises a programmable logic controller, also called a “PLC” system for “Programmable Logic Controller” in English, that is to say a control system for an industrial process comprising a man-machine interface. for supervision and a digital communication network. The PLC system can include several modular controllers which control the subsystems or control equipment of the installation. These devices are each configured to ensure at least one operation among: the acquisition of data from at least one measurement sensor, the control of at least one actuator connected to at least one flow controller unit, the regulation and the slaving of parameters, data transmission between the different equipment of the system.

The control unit 5 can thus comprise at least one of: a microcontroller, a microprocessor, a computer. The control unit 5 can be connected to the various control equipment of the installation, in particular to the flow regulating members 41, 42, to the sensor 8, and to communicate with said equipment by electrical, Ethernet, Modbus, etc. links. '' other connection methods and / or transmission of information, are possible for all or part of the equipment of the installation, for example by radiofrequency links, WIFI, Bluetooth, etc.

Firstly, the electronic logic 5 calculates a predetermined proportion of the flow rate D1 relative to a production flow rate DP and / or a predetermined proportion of the flow rate D2 relative to DP, i. e. predetermined D1 / DP and / or D2 / DP ratios, as a function of a target C1 content of the gas mixture in the doping gas and / or a target C2 content of the gas mixture in the carrier gas.

Preferably, the electronic logic 5 does not calculate the flow rate of the carrier gas D2 from a target content C2 of carrier gas but regulates D2 by deduction from D1. D2 then corresponds to DP from which D1 is subtracted. Preferably, the electronic logic 5 calculates a predetermined proportion of the flow rate D1 relative to DP from a target content C1 which is that of the minority gas in the mixture.

Note that for a ternary mixture for example, the adjustment of D1 and D2 can be done from respective target contents C1, C2, the third flow rate D3 setpoint in the third gas being deduced from the values of D1 and D2.

According to one possible implementation, the control unit 5 comprises a man-machine interface 300 comprising an input interface, for example a touch screen, allowing a user to input said at least one target content of the doping gas and / or carrier gas in the gas mixture. For example, the contents can be expressed as a volume percentage of the prime or carrier gas present in the gas mixture. More generally, the man-machine interface 300 can allow the user to give instructions to the control unit 5.

The flow rate regulators 41, 42 are instructed by the control unit 5 to regulate the flow of dopant gas and carrier gas to the respective setpoints D1, D2 determined from the target composition for the gas mixture. It is with these flow rates that the doping gas and the carrier gas enter the mixer device 3.

Typically, the mixing device 3 comprises a common mixing volume into which the inlet 32 and outlet 33 open and into which the mixture is homogenized. It is possible for example to use a mixer 3 of the static mixer type allowing continuous mixing of the fluids entering the mixer. This type of mixer generally comprises at least one interfering element, such a plate, a portion of pipe, an insert, capable of disturbing the flow of fluids, generating pressure drops and / or turbulence to promote the mixing of the fluids and its homogenization.

A mixture of dopant gas and carrier gas is therefore produced at the outlet 33 of the mixing device 3 with a production flow rate DP. The flow rates D1 and D2 are conditioned by the flow rate DP and by the desired contents C1, C2 of doping gas and carrier gas.

A problem which arises relates to the distribution of a gas mixture to a consumer unit 10 whose demand for the gas mixture is fluctuating. As a result, the rate of delivery of the gas mixture to point 10 will vary.

In order to adapt the flow of gas mixture produced at the outlet of the mixing device to the flow of gas consumed, the present invention proposes to connect the outlet 33 of the mixer 3 to the inlet of a buffer tank 7 via the outlet pipe 23. A distribution line 6 is fluidly connected to an outlet of the buffer tank 7 and makes it possible, in operation, to distribute the mixture to the consuming unit 10.

Note that the installation can include a vent line 25 fluidly connected to the buffer tank 7 with a vent 15 associated with a valve, useful in the event of overpressure, and to a valve controlling the passage of the mixture to a control unit. gas reprocessing. The valve makes it possible, during the start-up phases of distribution to the consumer unit, to purge the pipes of the installation and the buffer tank 7.

The distribution of the gas mixture to the consumer unit 10 therefore takes place from the buffer tank 7 with a DC consumption flow rate corresponding to the consumption of the mixture by the consumer unit 10. If the DC flow rate varies during the operation of the distribution installation, the production flow rate DP upstream of the buffer tank 7 may no longer correspond to the demand for mixing. The buffer tank 7, thanks to the additional volume that it provides on the fluid circuit, makes it possible to ensure distribution at the DC flow rate even if it does not correspond to the DP flow rate. In particular, if DP is greater than DC, the reservoir 7 prevents the gas mixture from being forced towards the distribution line and thus absorbs the overproduction. And if DP is less than DC, the buffer tank 7 forms a mixture reserve from which the user can draw, for example when consumption starts too much. quickly with a high consumption rate, which ensures distribution at the DC rate even in a situation of under-production.

In addition, the installation comprises a measurement sensor 8 which measures a physical quantity whose variation is representative of a variation in the DC consumption flow rate flowing in the distribution line 6 and provides a first measurement signal corresponding to the 'control unit 5. In particular, the first measurement signal can comprise several successive measurements carried out by the sensor 8. The unit 5 receives it and generates a first control signal which is transmitted to the flow regulating members 41, 42 of so as to adjust the first flow rate setpoint D1 and the second flow rate setpoint D2 in accordance with the first control signal.

The present invention thus makes it possible to recalculate the flow setpoints D1, D2 initially configured in order to adapt them to a variation in the DC consumption flow rate and therefore at the request of the user. The mixing device 3 produces a mixing flow rate, the control of which is associated with the flow rate consumed.

Note that in parallel, the control unit 5 continues to control the D1 / DP and D2 / DP ratios so that they comply with the doping gas and carrier gas contents desired for the gas mixture.

The method according to the invention advantageously implements a so-called start-up phase at the start of consumption of the mixture by the consuming unit, while no consumption was detected before. During this start-up phase, we go from a zero DP production flow rate to a production of a mixture of dopant gas and carrier gas with a predetermined DP production flow rate.

In practice, in the start-up phase, the user can start the production of the gas mixture with a predetermined flow rate DP which can be set at a minimum so-called starting value corresponding to a predetermined percentage of the maximum production flow rate that can be produced. This maximum production flow rate corresponds to the sum of a first maximum flow rate value and a second maximum flow rate value that the first and second regulating members 41, 42 are designed to distribute. Advantageously, the predetermined percentage is at least 25%, preferably at least 35% and more preferably at least 50% of the maximum production rate. This allows the use of the sensor that measures the flow in the flow regulators D1, D2 in its optimum and most precise operating range.

Note that during distribution to the consumer unit, the product gas mixture can be distributed to the vent 15, in particular in the case where the composition of the mixture does not comply with the target composition.

The user can optionally initially set a higher production flow rate than the expected DC consumption flow rate in order to fill the buffer tank 7 and constitute a mixture reserve there.

After the consumption start-up phase, a production regulation phase follows during which the production flow rate DP is adjusted as a function of the consumption flow rate DC. During the regulation phase, the control unit 5 monitors the DC consumption rate via the measurements received from the measuring sensor 8. If a change in the DC consumption rate is detected, the control unit 5 generates a first control signal for adapting the flow rates D1, D2 distributed upstream of the mixer in order to bring the flow rate DP in line with the modified flow rate DC.

Preferably, the measurement sensor 8 performs continuous or quasi-continuous measurements. Preferably, the control unit 5 is configured so that the generation of the first control signal and / or the transmission of the first control signal to the flow rate regulators only takes place at a predetermined time interval, in particular an interval of the order of 1 to 60 seconds. In other words, the flow setpoints are maintained during this time interval, without an adjustment of the setpoints being ordered by the control unit 5. This makes it possible to avoid a reaction of the installation following untimely fluctuations in the temperature. DC flow rate or to avoid generating too rapid variations of the DP flow rate which could give rise to operating errors.

Optionally, depending on the amplitude and / or the rate of variation of the DC flow rate, the control unit 5 can be configured to, at least temporarily, maintain the production flow rate DP. For example, if the DC consumption rate increases, the consuming unit 10 can draw from the buffer tank 7 to compensate for the underproduction of the mixer 3. If the DC consumption rate decreases, the buffer tank 7 can be filled to dampen the pressure. overproduction of mixer 3. Preferably, the control unit 5 is configured so as to stop the gas flows when the physical quantity measured by the sensor 8 is representative of a zero DC consumption flow rate. Thus, in the absence of demand, the installation does not produce a gas mixture. The control unit 5 can also be configured to stop the gas flows if the physical quantity measured by the sensor 8 is representative of a DC consumption flow rate is low, ie less than a given low flow threshold, in order to avoid overpressure in the buffer tank 7. The control unit 5 can also be configured to generate an alarm signal when the physical quantity measured by the sensor 8 is representative of a DC consumption flow rate greater than a flow rate threshold high given.

Advantageously, the installation according to the invention implements a first control loop of the first and second flow setpoints D1, D2 on the first measurement signal. By "control loop" is generally meant a control system of a process in which a controlling variable acts on a controlled variable, i. e. a quantity to be controlled, to bring it as quickly as possible to a setpoint value and to maintain it there. The basic principle of a servo-control is to measure, permanently, the difference between the real value of the quantity to be controlled and the set-point value that one wishes to reach, and to calculate the appropriate command to be applied to one or more. several actuators so as to reduce this gap as quickly as possible. This is also referred to as a closed-loop controlled system.

In the first control loop, the controlling variable is the physical quantity measured by the measuring sensor 8, the controlled variable is the production flow rate DP, via the setting of the flow rates D1 and D2 of doping gas and carrier gas. The setpoint is variable according to the consumption conditions of the mixture.

Besides the sensor 8, the first servo loop comprises a first comparator 11 A arranged within the control unit 5 and configured to generate at least a first error signal from the first measurement signal. The first error signal can be representative of a variation in the physical quantity measured. It is advantageously obtained by comparison with at least one measurement of said physical quantity taken at another time. In addition, the first loop comprises a first corrector 12A arranged within the control unit 5 and configured to generate the first control signal from the first error signal.

The first corrector 12A sends the control signal to actuators which control a movement, in response to the first control signal, of the first and second flow rate regulating members 41, 42 in respective positions in which the first flow setpoint D1 and the second flow setpoint D2 are adjusted in accordance with the first control signal. Typically, the actuators control the movement of moving parts within the regulators, which vary the flow rates D1, D2 sent to the mixing device 3 in a direction tending to reduce the difference between the flow rates DP and DC.

Preferably, the first corrector 12A is of the proportional, integral and derivative (PID) type, which makes it possible to improve the performance of a servo-control thanks to three combined actions: a proportional action, an integral action, a derivative action.

Preferably, and as mentioned above, the corrective action of the first servo loop is only applied to the setpoints D1, D2 at a predetermined time interval, preferably an interval between 1 and 60 s, preferably still of the order of 20 s, in order to avoid excessively rapid variations in the production flow rate which can create errors. This time interval can be a parameter of the first corrector 12A.

The first corrector 12A can include in particular a microprocessor, memory registers, programming instructions for processing the first error signal and for developing by numerical calculation the terms proportional, integral, and derivative of the control loop. These terms, which can be determined by calculation and / or experimentally, are combined to provide the control signal for the regulatory organs 41, 42. The term derived from D can optionally be zero.

FIG. 1 illustrates an embodiment in which the measurement signal is obtained by a flow sensor 8, also called a flow meter, arranged on the distribution line 6 so as to directly measure the DC consumption flow distributed to the consuming unit 10. The signals received and sent to the various elements of the installation are shown schematically by the dashed lines referenced "A". Typically, if the DC flow rate increases, the control signal orders an increase in the first and second flow rate setpoints D1, D2 and a decrease in the first and second flow rate set points D1, D2 if the DC flow rate decreases.

Note that in the context of the invention, each of the first and second flow regulating members 41, 42 can move between a closed position in which the first flow setpoint D1 or the second flow setpoint D2 is zero and a fully position. open in which the first flow rate instruction D1 or the second flow rate instruction D2 respectively have a first maximum flow rate value or a second maximum flow rate value.

The first and second flow regulating members 41, 42 can optionally occupy at least one intermediate position between the closed position and the open position. Preferably, said intermediate position corresponding to a first flow rate setpoint D1 or a second flow rate setpoint D2 greater than or equal to a first minimum flow rate value or a second minimum flow rate value. Preferably, the first minimum flow rate value and / or the second minimum flow rate value is equal to at least 25%, more preferably at least 35%, or even at least 50%, of the respective first or second maximum value. This makes it possible to work on flow ranges where the precision of the regulating members 41, 42, more precisely the precision of the flow sensors used in the regulating members, is better.

Optionally, these positions can be predefined, to increase in an incremental and controlled way the flow rates in the desired range, which allows better control of the mixing precision, thanks to the first servo loop.

According to an alternative embodiment, the installation uses a pressure sensor 8 measuring the pressure prevailing in the buffer tank 7 as a physical quantity representative of the DC consumption rate. The DC consumption flow fluctuations are thus determined indirectly, via the determination of pressure fluctuations in the buffer tank 7. The representation of FIG. 1 remains applicable except that the measurement signal is produced by the sensor 8 connected to the tank. buffer and not by sensor 8 connected to line 6.

Note that the installation according to the invention can include two sensors 8, one for flow and the other for pressure. These sensors are as described above and each produce a respective first measurement signal. As a function of predetermined selection criteria, the control unit 5 is configured to generate the first control signal from the measurement signal coming from one or the other of the sensors 8. Preferably, the control unit 5 chooses to use the first measurement signal originating from that of the two measurement sensors 8 which measures a physical magnitude value representative of the highest flow rate.

In practice, if the consumption rate DC increases, the production rate DP produced at the outlet of the mixing device 3 will begin to become insufficient. The consuming installation 10 will draw from the buffer tank 7 to compensate for the under-production of the mixer 3, resulting in a decrease in the pressure in the tank 7.

The pressure sensor 8 sends the first measurement signal to the first comparator 11 A which generates a first error signal corresponding to the pressure drop information and transmits it to the first corrector 12A so that it calculates a first control signal applied to the first and second flow rate regulating members 41, 42 so that the first and second flow rate setpoints D1, D2 increase by an appropriate factor, which can be determined by the first regulation loop.

According to one embodiment, the first comparator 11A is configured to generate at least a first error signal from a comparison of the first measurement signal with at least one parameter chosen from: a low pressure threshold, a threshold of high pressure. These thresholds can be adjusted according to the operating conditions, the characteristics of the installation, etc. When the pressure in the buffer tank 7 reaches the low pressure threshold, the first corrector commands the flow regulating members to regulate the flow. dopant gas and carrier gas according to the flow rate instructions D1, D2 given.

This operating mode can be implemented during the regulation phases as well as during the consumption start-up phases. In the case of a start-up phase, as soon as the pressure in the buffer tank 7 reaches the low pressure threshold, the flow regulating members are commanded to regulate the flow of dopant gas and carrier gas so as to produce the mixture. gas with the DP flow rate set at the start value. In particular, the flow setpoints D1, D2 can correspond respectively to the first minimum flow rate value and the second minimum flow rate value. The flow rate regulators 41, 42 each start to produce minimum flow rates leading to a DP flow rate equal to the starting value until the high pressure threshold in the buffer tank 7 is reached.

According to one possibility, if the pressure in the reservoir 7 does not increase sufficiently, in particular if the high pressure threshold is not reached, or if the pressure does not increase quickly enough, the flow setpoints D1, D2 are increased by following a regulation scheme by the first corrector 12A, preferably of the PID type, in which the increase in flow rates is a function of the drop in pressure.

If the pressure in the reservoir 7 reaches the high pressure threshold, the flow rate regulating members 41, 42 can be moved to their respective closed positions in which the flow rates D1, D2 are impaired.

FIG. 2 shows schematically an example of the effect of a first servo loop with a first PID type corrector in which the production rate DP, corresponding to the sum of D1 and D2, is corrected as a function of the variation of the pressure P7 in the buffer tank 7. The maximum production flow DP of the installation, corresponding to the sum of the first and second maximum flow values, is set at 100 sL / min (standard liter per minute), ie 6 Nm3 / h (normo cubic meter per hour). The installation's minimum DP production flow rate, corresponding to the sum of the first and second minimum flow rate values, is set at 25 sL / min (standard liter per minute), i.e. 1.5 Nm3 / h. The high and low pressure thresholds are set at 4 bar and 3.8 bar respectively.

FIG. 2 schematically represents various scenarios which may be encountered during the operation of the installation. If DP = DC, the pressure remains stable at 4 bar (gray arrow at the bottom right of figure 2). Subsequently, assuming DOO but DP = 0, the pressure in the buffer tank will drop to 3.8 bar (moving to the left along the gray arrow). This pressure is the starting pressure of the flow regulators. The DP flow rate is at its minimum start-up value, ie 25 sL / min As soon as the control unit has commanded the flow regulators to produce a DP flow <DC, the pressure will drop until a temperature is reached. DC flow rate equal to the maximum DP flow rate of the installation, ie 100 sL / min (movement from bottom to top along the gray arrows). As soon as DC decreases, i.e. DP> DC, the buffer tank begins to fill and the pressure increases from 3.5 bar to 4 bar (following the arrows with black lines). 4 bar is the stop pressure for filling the buffer tank.

An example of what happens in practice is shown in Figure 3 showing the time evolution of the pressure in the buffer tank (dashed curve) and the production flow DP (solid line). At the start of the graph (zone A) if there is no drop in pressure, the flow setpoint remains at 0. As soon as the pressure drops (zone B), flow setpoints are given to the flow regulators D1 and D2, which are incremented at a regular interval if the pressure does not stabilize. As soon as the pressure has stabilized, filling the buffer tank (zone C) is stopped. If the pressure drops again (zone D), the setpoints of the flow regulators will be adjusted to the desired values in order to allow the DC consumption to be predicted and to keep the pressure of the buffer tank stable.

Note that the normo cubic meter is a unit of measurement of quantity of gas which corresponds to the content of a volume of one cubic meter, for a gas found in normal conditions of temperature and pressure (0 or 15 or more rarely 20 ° C according to the standards and 1 atm, i.e. 101 325 Pa). For a pure gas, a normal cubic meter corresponds to approximately 44.6 moles of gas.

Note that the buffer tank advantageously has an internal volume equal to at least half of the maximum DP production flow rate DP of the installation.

DPmax

Minimum buffer volume = -; -

2

Respecting this minimum internal volume makes it possible to absorb the pressure variations linked to the untimely nature of DC. The buffer tank can have an internal volume of at least 1 L, or even at least 50 L, or even 1000 L or more. Preferably, the internal volume of the buffer tank will be between 50 and 400 L. The tank can be formed from a single tank or from several tanks fluidly connected to one another, the internal volume of the buffer tank then being understood as the sum of tank volumes.

In an advantageous embodiment, visible in FIG. 1, the installation may further comprise a first analysis unit 13 configured to analyze at least one content of the doping gas and / or the carrier gas of the distributed gas mixture. via the supply line 6. This makes it possible in particular, during the start-up phase of the installation, to condition the distribution of the gas mixture to the conformity of the measured contents with the target contents. A tolerance of the order of 0.01 to 5% (relative%) with respect to the target levels C1, C2 can be set. If the mixture produced does not conform, production may possibly be stopped. Preferably, the first analysis unit 13 is configured to analyze the dopant gas content, which may in particular be the minority gas in the gas mixture.

Advantageously, the installation comprises a first sampling duct 36 connecting the first analysis unit 13 to the supply line 6 at a first sampling point 36a. Part of the mixture flowing in the supply line 6 from the reservoir 7 is thus taken by the first sampling duct 36 to be analyzed in the first analysis unit 13. After passing through the first analysis unit 13 , the mixture withdrawn returns to the supply line 6 via a first return line 37 connected to the supply line 6 at a first return point 37a which is located downstream from the first withdrawal point 36a on the line of Feed 6. Since the gas mixture is a high-precision, high-value-added doping gas, this recirculation scheme avoids the rejection and loss of the mixture. In addition, there is no need for any reprocessing of the rejected mixture, which would be costly and complex for the user given the nature of the gases used.

The installation further comprises at least one regulator 51 mounted on the supply line 6 between the first sampling point 36a and the first return point 37a. The regulator operates as a downstream pressure reducer and makes it possible to ensure the pressure differential necessary for the flow of the gas mixture in the first sampling and return conduits 36, 37. In addition, the regulator 51 is configured to regulate the flow. pressure of the gas mixture supplied to the silicon wafer doping unit 10. This ensures the stability of the pressure at the point of use of the mixture in order to meet the precision requirements of a silicon doping unit. and stability of the mixture parameters. In particular, the regulator 51 can be mounted in series on the supply line 6.

The installation according to the invention can also include a second analysis unit 14 arranged upstream of the buffer tank 7 so as to measure at least one content of the doping gas and / or the carrier gas of the gas mixture produced by the device. mixer 3. Depending on the case, the invention may comprise one and / or the other of the first 13 and second 14 analysis units. The second analysis unit 14 is configured to consequently supply at least one second measurement signal to the control unit 5, which generates a second control signal from the second measurement signal. The second control signal is used to control one and / or the other of the flow rate regulating members 41, 42 so as to adjust one and / or the other of the proportions of the first flow rate set point D1 and of the second flow rate setpoint D2 with respect to the production flow rate DP so that the effective composition of the gas mixture leaving the mixing device 3 approaches the target composition at levels C1, C2 (C2 being preferably deduced from C1 and not measured ). The signals received and sent to the various elements of the installation as part of the control of the composition of the mixture are shown diagrammatically by the dashed lines “B”.

This control of the contents of the mixture produced by the mixing device makes it possible to compensate for any errors between the flow rates actually set by the flow regulating members 41, 42 and the flow setpoints D1, D2 which are applied to them. The arrangement of a sampling point located between the outlet of the mixing device and the inlet of the buffer tank 7 makes it possible to detect and react more quickly to any variations in content, thus avoiding the risk of consuming an incorrect mixture. .

Advantageously, the installation comprises a second sampling duct 34 connecting the second analysis unit 14 to the outlet pipe 23 at a second sampling point 34a and a second return duct 35 connecting the second analysis unit 14 to the outlet pipe 23 at a second return point 35a, the return point 35a being located downstream of the first sampling point 34a on the outlet pipe 23. As already explained, the gas mixture being a high precision doping gas and high added value, this recirculation scheme avoids the rejection and loss of the mixture. In addition, there is no need for any reprocessing of the rejected mixture, which would be costly and complex for the user given the nature of the gases used.

The installation further comprises at least one overflow device 52 mounted on the outlet pipe 23, between the second sampling point 34a and the second return point 35a. When the pressure varies upstream, the regulator then changes the flow rate in the bypass line so that its inlet pressure remains constant and a constant flow passes through the outlet line 23. In fact, the regulator 52 comprises a member which closes when the upstream pressure is greater than a predetermined threshold. The regulator 52 opens and becomes passing at a determined flow rate when the upstream pressure is below this threshold, or as a function of a pressure differential between the upstream and downstream ends of the regulator.

According to one embodiment, the regulator may include a chamber mounted as a bypass, a valve controlled by a control membrane. This membrane is balanced on the one hand by a calibrated spring intended to close and open a conduit connected to the gas circuit and on the other hand by the pressure to be stabilized upstream.

The regulator 52 performs several functions. It operates as an upstream pressure regulator, that is to say it is configured to regulate the pressure of the gas mixture in the gas circuit upstream of said overflow device 52, in particular at outlet 33, in mixer 3, at the inlet 31 of the mixer, at the level of the regulators 41, 42.

During the production regulation phases during which a DP flow rate is produced and adjusted as a function of the DC flow rate, the buffer tank 7 is filled and the pressure in the tank 7 varies according to the variations in consumption. These pressure fluctuations are also found at the inlet 31, in the pipes 21, 22 in communication with the reservoir, which can distort and / or disturb the flow measurements carried out by the flow regulating members 41, 42. The use of the regulator 52 makes it possible to maintain the upstream pressure constant, while the downstream pressure can fluctuate. This greatly improves the precision and the stability of the composition of the doping mixture.

Furthermore, when consumption ceases, the pressure in the reservoir 7 tends to increase. As soon as the flow DP stops, the regulator 52 encloses the mixture in the upstream circuit, which makes it possible to maintain it at the desired pressure when the installation is stopped. On start-up, when the mixer 3 begins to produce the mixture at the flow rate DP, the regulator makes it possible to reduce the time necessary for the flow regulators 41 42 to reach their set points, that is to say the starting time of the flow regulating members 41, 42. Typically, it has been possible to obtain response time of the regulators 41, 42 of less than 1 second, or even less than a few milliseconds.

The regulator 52 also makes it possible to ensure the pressure differential necessary for the flow of the gas mixture in the first sampling and return conduits 36, 37.

Note that the second sampling duct 34 taking the mixture and leading it into the analysis unit 14 advantageously has the shortest possible length so that the analyzer provides a very precise response in real or near real time. Preferably, the pipe is such that the time lag between the moment when the mixture is taken at its sampling point and the moment when the analysis unit gives its measurement is minimal, typically less than 30 seconds, in particular between 1 and 30 seconds.

Preferably, the second control signal is produced from a second error signal containing at least one item of information on the difference between a measured content and a target content, for the doping gas or the carrier gas. Preferably, only the content of the doping gas is measured and compared to its target value, the doping gas being the minor gas of the mixture. This difference can be expressed in particular as: where M1 is the measured content for the dopant gas. The relative deviation DC1 can be used as a correction factor for the first flow setpoint D1.

Let us consider the example of an installation configured to produce a two-gas mixture with a production flow rate DP at the outlet of the mixing device 3 of 100 sL / min. The desired gas mixture is a mixture formed of the doping gas with a target C1 content of 0.5% and the carrier gas for the rest, therefore with a C2 content of 99.5% (% by volume). A premix comprising a pure doping substance diluted to 30% by volume in a carrier gas is used for the flow rate D1. A first flow setpoint D1 of 1.667 sL / min (0.1 Nm3 / h), corresponding to a proportion of 1.667% with respect to DP, and a second setpoint D2 of 98.333 sL / min (5.1 Nm3 / h) corresponding to a proportion of 98.333% relative to DP, are therefore applied to the respective flow regulating members 41, 42. We assume a precision of adjustment of the members 41, 42 of plus or minus 1%. An error of -1% on D1 and +1% on D2 leads to an actual flow rate of doping gas equal to 1.650 sL / min, to an actual flow rate of carrier gas equal to 99.316 sL / min and to a flow rate of actual production of 100.967 sL / min. A doping gas content of 0.49% is measured at the outlet of the mixing device 3, corresponding to a DC1 deviation of −1.95% (relative%) with respect to the target content C1. The control unit 5 generates a second control signal commanding the flow rate regulating members 41, 42 to adjust the flow rates D1 and D2 with respect to DP so as to compensate for this difference. The first setpoint D1 is therefore adjusted to D1 = 1, 682 sL / min.

Preferably, only the first setpoint D1 is adjusted as a function of the second measurement signal, the control unit 5 controlling the maintenance of D2. It being understood that it is conceivable that D2 also is adjusted in response to the second control signal. In the example above, D2 would be adjusted to 97.4 sL / min. Note that the correction can also be made by applying a correction factor to at least one of the target levels previously recorded in the control unit 5, in the above example a correction by a factor equal to 0.78%, which has the effect of adjusting D1 to 1.682 sL / min accordingly.

Optionally, the installation can include an alarm configured to emit an alarm signal if the first analysis unit and / or the second analysis unit detects levels outside the planned tolerance ranges.

The first analysis unit 13 and / or the second analysis unit 14 can be chosen in particular from the following types of detectors: a thermal conductivity detector, a paramagnetic alternating pressure detector, a catalytic adsorption detector, a detector with non-dispersive infrared absorption, an infrared spectrometer, an acoustic or photoacoustic wave propagation gas concentration analyzer. The type of analysis unit can be adapted according to the nature of the gases to be analyzed. Optionally, the first 13 and second 14 analysis units can be swapped.

According to one embodiment, the installation may include a second control loop of the respective proportions of the first flow rate setpoint D1 and / or of the second flow rate setpoint D2 relative to the production flow rate DP on the second measurement signal. supplied by the second analysis unit 14. In the second control loop, the controlling quantities are the content (s) measured by the second analysis unit 14, the controlled quantities are one and / or the other of the proportions D1 / DP, D2 / DP. The setpoint is variable depending on the actual content (s) measured.

The second loop comprises a second comparator 11 B arranged within the control unit 5 and configured to generate at least a second error signal from a comparison of the second measurement signal with at least one parameter chosen from: the target content C1 in the doping gas, the target content C2 in the carrier gas. A second corrector 12B is arranged within the control unit 5, in particular of the PID type, and configured to generate the second control signal from the second error signal. In response to the second control signal, the actuators of the first and second flow rate regulators 41, 42 control the movement of the first and second flow rate regulators 41, 42 into respective positions in which the proportions of D1 and / or D2 by compared to DP conform to the second control signal. Preferably, only the proportion of D1 is adjusted, the regulation loop ordering D2 to remain fixed.

It should be noted that the first comparator and the second comparator can possibly form the same entity configured to receive as input data both the measurements of the sensor 8 and of the second analysis unit 14 and to produce the appropriate error signals as output. . It is the same for the first and second correctors.

The installation according to the invention can be used for the distribution of gas mixtures used in different industries such as semiconductor, photovoltaic, LED, flat screen industries or any other industry such as mining, pharmaceutical, space industries. or aeronautics.

Preferably, the installation comprises at least one gas cabinet (in English "gas cabinet") in which are installed at least the control unit 5, the mixing device 3, the flow regulating members, the measurement sensor 8. , the buffer tank 7. The sources of doping gas and carrier gas can be located inside or outside the cabinet. Preferably, the sources are located outside the cabinet so that the latter retains a reasonable size. Preferably, the unit of control 5 is arranged outside the cabinet, either by being fixed to one of the walls of the cabinet, or positioned at a distance from the cabinet.

The gas cabinet may include a housing with a back wall, sidewalls, a front wall, a bottom, and a ceiling. In the housing, one or more buffer reservoirs are provided which stand on the bottom and can be fixed in the housing in a manner known in the state of the art. A gas line system is arranged in said housing, preferably against the bottom of the cabinet. The cabinet may include means for monitoring and / or maintaining the gas pipe system such as valves, pressure reducers, pressure measuring devices, etc. making it possible to carry out operations such as gas distribution, the opening or closing of certain conduits or portions of conduits, gas pressure management, carrying out purge cycles, leak tests, etc.

The housing includes gas inlet openings for supplying dopant gas and carrier gas and a gas outlet opening for dispensing the gas mixture. The distribution line 6 is connected to the outlet opening. In operation, the gas cabinet is connected to the consumption unit by the distribution line 6. Other gas inlets may be provided, in particular for a purging gas or a gas which creates a vacuum by the venturi effect, and a standard gas for calibrating the analyzers.

The installation according to the invention can in particular be used to produce gas mixtures having the following compositions:

- 2% of AsH3 in Ar,

- 1 to 10% of AsH3 in He, in particular contents of 1%, 2% or 10% of AsH3 in He,

- 1 to 20% of AsH3 in H2, in particular contents of 1%, 3%, 4%, 5%, 7%, 10%, 15% or 20% of AsH3 in H2,

- 1 to 10% of AsH3 in N2, in particular contents of 1%, 2%, 5% or 10% of AsH3 in N2,

- 1 to 10% of B2H6 in Ar, in particular contents of 1%, 2%, 3%, 4%, 5% or 10% of B2H6 in Ar,

- 1 to 10% of B2H6 in H2, in particular 1% or 10% of B2H6 in H2, - 1 to 10% of B2H6 in N2, in particular 1%, 2%, 3%, 4%, 5% or 10% of B2H6 in N2,

- 1 to 15% of PH3 in Ar, in particular 1%, 2%, 5%, 10% or 15% of PH3 in Ar,

- 1 to 10% of PH3 in He, in particular 1%, 2% or 10% of PH3 in He,

- 1 to 15% of PH3 in H2, in particular 1%, 5%, 10% or 15% of PH3 in H2,

- 1 to 15% of PH3 in N2, in particular 1%, 2%, 3%, 4%, 5%, 10% or 15% of

PH3 in N2.

Preferably, the target contents C1 of the doping gas are between 0.0001 and 50%, preferably between 0.1 and 30%, the remainder being the carrier gas.

In order to demonstrate the efficiency of an installation according to the invention, a mixture comprising diborane (B2H6) as doping gas in hydrogen as gas was produced and distributed on site. carrier. The doping gas consisted of a 20% by volume diborane premix diluted in hydrogen. The installation included a first PID-type feedback loop as described above and a second feedback loop.

In a first test, corresponding to FIG. 4, the mixing was carried out at levels of B2H6 increased in stages, in order to show the precision and the resolution that can be obtained in the doping gas content. An accuracy of 0.005% (absolute%) could be achieved on the B2H6 content.

In a second test, corresponding to FIG. 5, a mixture was produced with a target C1 B2H6 content of 0.5% (% by volume) and this content was measured during fluctuations in consumption. of the doping unit. Fig. 5 shows a recording of the flow rate of the DC gas mixture delivered by the distribution line with the content of B2H6 measured during this recording.

A variable DC gas mixture flow rate between typically 0 and 30 sL / min could be produced with a stability of the B2H6 content characterized by a relative standard deviation of the order of 0.008% (% absolute) or 80 ppm in absolute, or 1.6% in relative. The measured content was 0.494% on average. The horizontal lines indicate the minimum and maximum values reached by the B2H6 content during recording. It should be noted that the present description describes a gas mixture with two constituents but that it can be transposed to any mixture having a greater number of constituents. For example, in the case of a ternary gas mixture, three sources each distribute a doping gas, a carrier gas, a third gas. Flow regulating members 41, 42, 43 are instructed by the control unit 5 to regulate the flow of the first, carrier gas and third gas to respective flow setpoints D1, D2, D3. The mixer device is configured to distribute a mixture of flow rate DP equal to the sum of D1, D2, D3. The proportions of first, carrier gas and third gas relative to DP are determined as a function of at least two among three target contents C1, C2, C3 of the gas mixture of the dopant gas, the carrier gas and the third gas respectively. All or part of the characteristics already described for a two-gas mixture can be transposed to this three or more gas mixture.

Claims

1. Installation for distributing a gas mixture suitable for and intended for use in a silicon wafer doping unit, said installation comprising:

- a source of a doping gas (1),

- a source of a carrier gas (2),

- a mixing device (3) fluidly connected to the doping gas container (1) and to the carrier gas source (2), said mixing device (3) being configured to produce at an outlet (33) a gas mixture comprising the dopant gas and carrier gas,

- a first flow regulator member (41) and a second flow regulator member (42) configured to respectively regulate the flow rate of the doping gas (1) and the flow rate of the carrier gas (2) flowing towards the mixing device (3) ) according to a first flow setpoint (D1) and a second flow setpoint (D2) defining, in operation, a production flow rate (DP) of the gas mixture at the outlet (33) of the mixing device (3),

- a control unit (5) configured to control the first and second flow regulator members (41, 42) so as to adjust the first flow setpoint (D1) and the second flow setpoint (D2) according to respective proportions by relative to the production flow rate (DP), said respective proportions being determined as a function of at least one target content (C1, C2) of the gas mixture of the doping gas (1) and / or the carrier gas (2),

- a buffer tank (7) connected by an outlet pipe (23) to the outlet (33) of the mixing device (3) on the one hand and to a distribution line (6) on the other hand, the distribution line (6) being configured to distribute the gas mixture to a silicon wafer doping unit (10) with a consumption rate (DC) representative of a variable consumption of the gas mixture,

- at least one measurement sensor (8) configured to measure a physical quantity whose variation is representative of a variation in the consumption flow rate (DC) distributed by the distribution line (6) and to provide a first measurement signal of said physical quantity, the control unit (5) being connected to the measurement sensor (8) and configured to generate a first control signal from the first measurement signal, the flow regulating members (41, 42) being configured to adjust the first flow setpoint (D1) and the second flow setpoint (D2) in response to said first control signal.

2. Installation according to claim 1, characterized in that it comprises a first analysis unit (13) arranged downstream of the buffer tank (7) and configured to analyze at least one respective content of the dopant gas (1) and / or the carrier gas (2) of the gas mixture distributed by the supply line (6).

3. Installation according to claim 2, characterized in that it comprises a first sampling duct (36) connecting the first analysis unit (13) to the supply line (6) at a first sampling point (36a). ) and a first return line (37) connecting the first analysis unit (13) to the supply line (6) at a first return point (37a), the return point (37a) being located downstream from the first sampling point (36a) on the supply line (6), a regulator (51) being mounted on the supply line (6) between the first sampling point (36a) and the first return point ( 37a), preferably the pressure reducer (51) is mounted upstream of the measurement sensor (8).

4. Installation according to one of the preceding claims, characterized in that it comprises a second analysis unit (14) configured to measure at least one content of the doping gas (1) and / or the carrier gas (2) of the gas mixture produced at the first outlet (33) of the mixing device (3) and consequently to supply at least one second measurement signal, the control unit (5) being connected to the second analysis unit (14 ) and configured to generate a second control signal from the second measurement signal and to modify the proportion of the first flow setpoint (D1) and / or the proportion of the second flow setpoint (D2) with respect to the flow rate of production (DP) in response to said second control signal.

5. Installation according to claim 4, characterized in that it comprises a second sampling duct (34) connecting the second analysis unit (14) to the outlet pipe (23) at a second sampling point (34a) and a second return duct (35) connecting the second analysis unit (14) to the outlet (23) at a second return point (35a), the return point (35a) being located downstream of the first sampling point (34a) on the outlet pipe (23), a regulator (52) being mounted on the outlet pipe (23), between the second sampling point (131a) and the second return point (132a).

6. Installation according to one of the preceding claims, characterized in that it is configured to distribute a mixture having a dopant gas content (1) between 0.0001 and 50%, preferably between 0.05 and 30% (% by volume).

7. Installation according to one of the preceding claims, characterized in that the source of doping gas (1) contains germanium tetrahydride (GeH4), phosphine (PH3), arsine (AsH3) and / or diborane (B2H6) and / or the carrier gas source (2) contains hydrogen (H2), nitrogen (N2) and / or argon (Ar).

8. Installation according to one of the preceding claims, characterized in that the source of doping gas (1) contains a gas premix formed of doping gas (1) and carrier gas (2).

9. Installation according to one of the preceding claims, characterized in that it comprises a first control loop of the first and second flow setpoints (D1, D2) on the first measurement signal supplied by the measurement sensor (8 ), said first loop comprising:

- a first comparator (11 A) arranged within the control unit (5) and configured to generate at least a first error signal from the first measurement signal,

- a first corrector (12A) arranged within the control unit (5), in particular of the proportional, integral and derivative (PID) type, and configured to generate the first control signal from the first error signal ,

- actuators of the first and second flow regulator members (41, 42) connected to the first corrector (12A) configured to receive the first control signal and move the first and second flow regulator members (41, 42) into respective positions in which the first flow setpoint (D1) and the second flow setpoint (D2) comply with the first control signal.

10. Installation according to one of the preceding claims, characterized in that it comprises a second control loop of the respective proportions of the first flow setpoint (D1) and / or of the second flow setpoint (D2) relative to at the production flow rate (DP) on the second measurement signal supplied by the second analysis unit (14), the second loop comprising:

- a second comparator (11 B) arranged within the control unit (5) and configured to generate at least one second error signal from a comparison of the second measurement signal with at least one parameter chosen from among : a target content (C1) in the doping gas (1), a target content (C2) in the carrier gas (2),

- a second corrector (12B) arranged within the control unit (5), in particular of the proportional, integral and derivative (PID) type, and configured to generate the second control signal from the second error signal ,

- the actuators of the first and / or second flow regulator members (41, 42) connected to the second corrector (12B) and configured to move the first and / or second flow regulator members (41, 42) into respective positions in which the proportions of the first flow setpoint (D1) and / or of the second flow setpoint (D2) with respect to the production flow (DP) conform to the second control signal.

11. Installation according to one of the preceding claims, characterized in that the measurement sensor (8) comprises a flow sensor or flowmeter configured to measure the consumption flow (DC).

12. Installation according to claim 11, characterized in that the first comparator (11 A) is configured to generate at least a first error signal representative of a variation in the consumption rate (DC) and the first corrector (12A). is configured to generate a first control signal controlling a movement of the first and second flow regulating members (41, 42) so that the first and second flow setpoints (D1, D2) vary in the same direction as that of the variation flow rate (DC).

13. Installation according to one of the preceding claims, characterized in that the measurement sensor (8) comprises a pressure sensor configured to measure the pressure prevailing in the buffer tank (7).

14. Installation according to claim 13, characterized in that the first comparator (11A) is configured to produce a first error signal representative of a variation in the pressure in the buffer tank (7) and the first corrector (12A). is configured to generate at least a first control signal controlling a movement of the first and second flow regulating members (41, 42) so that the first and second flow setpoints (D1, D2) vary in a direction opposite to that of the pressure variation.

15. Assembly comprising a unit for doping silicon wafers comprising a furnace provided with an enclosure associated with heating means and a support arranged in said enclosure on which the wafers are installed, the furnace comprising introduction means. of a mixture of doping gas (1) and carrier gas (2) in the enclosure, characterized in that it further comprises an installation according to one of claims 1 to 14, said introduction means being connected fluidly to the supply line (6) of said installation.