FR3115485A1 - Block slicing process - Google Patents

Abstract

The cutting process makes it possible to cut into slices a block (100) integral with a support (1002). The cutting method includes a step of cutting the block (100) by a sheet (1001) formed by at least one cutting wire (1004). The cutting step comprises a relative rapprochement between the sheet (1001) and the support (1002) implemented while the cutting wire (1004) is scrolling, and successive phases each comprising: - a first period for which the rapprochement is carried out at a first closing speed, - a second period for which the closing is carried out at a second closing speed, the second closing speed being strictly lower than the first closing speed and the second closing speed being such that the cutting of the block (100) is carried out according to a ductile mode during the second period. Figure to be published with abstract: Fig. 1

Description

Block slicing process

Technical field of the invention

The technical field of the invention relates to the cutting, also called sawing, of a block into slices. Preferably, the wafers are silicon wafers intended to be used for photovoltaic applications in particular to form photovoltaic cells.

State of the art

The sawing of thin slices of silicon, or other hard substances such as sapphire, can be carried out in a block using a cutting wire in particular with bonded abrasives (also simply called wire with bonded abrasives). These bonded abrasives are in the form of grains bonded to a backing wire that comprises the cutting wire. These grains are generally diamonds, we then speak of diamond wire to designate the cutting wire comprising these diamonds. The cut is made using a longitudinal movement of the cutting wire, most often alternating (periodic reversal of direction), on which the block to be sawn exerts, for example, a supporting force. This cutting operation is carried out in a cutting machine where the cutting wire can be wound several hundred times around rotary cylindrical guide members with grooves, these grooves determining the spacing between turns of the winding of the cutting wire. cutting around the guide members. This winding results in the formation of a sheet which allows cutting by penetration into the block. The opposite longitudinal ends of the cutting wire are secured to two coils allowing the cutting wire to scroll back and forth from which it follows that the sheet can cut the block by sawing when a vertical displacement is applied to the block.

The block of silicon to be cut is generally fixed to a moving part of the cutting machine, this moving part being commonly called a table or support. A sacrificial material is used to mechanically fix the block of silicon to the moving part, for example by bonding. Thus, while the cutting wire scrolls due to its round trips, the movable part can be pushed downwards so as to cause the block of silicon to pass through the sheet, the sheet remaining in a fixed position and being arranged under the moving part. For a block of silicon with a section, for example square, of 156 mm by 156 mm, the useful travel of the mobile part is of the order of 165 mm so that the sheet ends its travel in the sacrificial material on which the silicon block is glued. It is then sufficient to peel off the slices of the sacrificial material to recover them. The table descent time defines the duration of the cut. It is customary to drive the block of silicon through the sheet as quickly as possible, in order to reduce manufacturing costs and amortize the price of the cutting machines as soon as possible. The cutting time is also linked to the maximum scrolling speed of the cutting wire which depends on the cutting machine: for a scrolling of the cutting wire at a speed of 30 m/s, the cutting of the block of silicon whose section is given above can reach 120 minutes or more depending on the moving speed of the table. The cutting wire speed is limited by the capabilities of the cutting machine.

According to a known cutting process, and with the aim of increasing the rate of production of wafers from the block of silicon, that is to say to cut a maximum of silicon wafers in a minimum of time, the manufacturers of Silicon wafers tend to maximize both table travel speed and cutting wire scroll speed. A high running speed of the cutting wire will tend to reduce the vertical force exerted on the cutting wire, which is preferable, in particular for the mechanical strength of the cutting wire (that is to say to limit wear cutting wire).

To avoid the breakage of the cutting wire and to obtain a good quality of silicon wafer defined by a maximum thickness difference strictly less than 20 μm and determined between the two opposite faces of the corresponding silicon wafer, the deformation of the cutting wire (that is to say the maximum deflection of the cutting wire in the center of the block during cutting) generally does not exceed ten millimeters. For this, apart from during the first minutes and the last minutes of a cut, corresponding respectively to the entry of the sheet into the block, then to its exit from the block and its entry into the sacrificial material, the speed of movement the table is usually held constant or nearly constant during cutting.

Conventionally, for obvious reasons of limiting the manufacturing cost of silicon wafers, there is a tendency to minimize the consumption of cutting wire. For example, from a 50 km cutting wire initially wound mainly on one of the reels, it is possible to cut 50,000 slices, i.e. an average cutting wire consumption of 1 m per 156 mm slice per 156mm. During the cutting, the cutting wire goes back and forth over a few hundred meters, with a slight difference between the length of the outward and the return, for example for 600 m of cutting wire unwinding from the one of the spools, only 594 m of cutting wire is rewound on this spool. The ratio between the length of the rewound cutting wire and the length of the unwound cutting wire, which here is 99%, is commonly called the return rate or “back and forth” rate in English.

This cutting process leads to obtaining slices with particular surface conditions. The periodic back and forth movements of the cutting wire, as the table makes its regular downward movement, results in the formation of slices with surfaces exhibiting periodic undulations. Visually, such ripples are associated with alternating light and dark streaks. These streaks are more or less deep depending on the direction of measurement of the thickness of the slice. The amplitude of these undulations measured on a wafer with a mechanical profilometer moving parallel to the cutting direction of the wafer is generally between 1 μm and 6 μm, with maximums that can reach 20 μm. The period of these undulations is close to 1 mm to 2 mm, and depends on the length of cutting wire running during each return trip and on the speed of movement of the table.

Furthermore, according to this cutting process, the passage of the diamonds from the diamond wire to the surface of the silicon leads to the formation on the surface of the slices cut of grooves which may appear smooth or have chippings which correspond to surface defects. This phenomenon, for example described in the document "Wire sawing technology: A state-of-the-art review" by Hao Wu and published in Precision Engineering 43 (2016) 1-9, corresponds to an alternation of plastic deformations of silicon ( ductile abrasion) and peeling of the surface (brittle abrasion). In fact, the diamond(s) slide on the surface of the silicon and create a smooth groove as long as the silicon deforms plastically, then when the pressure in front of the diamond is too great, the stress on the silicon causes the tearing off a piece in order to release this constraint.

The tearing of pieces of silicon during cutting can be associated with the propagation of cracks under the surface of the cut wafers. The area thus affected, in which these cracks are present, is called subsurface damage or subsurface damage, and is also known by the acronym SSD for "subsurface damage" in English. Such subsurface damage can impact the mechanical properties of the wafers and generate defects in the breasts of photovoltaic cells subsequently fabricated using the wafers. Such defects can be an inhomogeneity in the surface texturing, defects in the deposition of thin layers or defects in electrical contacts after screen printing.

Thus, there is a need to reduce surface defects and/or sub-surface damage for slices obtained by cutting a block.

Object of the invention

The object of the invention is to improve the cutting of slices in a block, in particular by making it possible to reduce the surface defects and/or the sub-surface damage of the cut slices in order in particular to improve the mechanical properties of the slices by example forming rough-sawn silicon wafers.

To this end, the invention relates to a method for cutting a block into slices, the block being integral with a support, the cutting method comprising a step of cutting the block by a layer formed by at least one cutting, the cutting step comprising a relative bringing together between the sheet and the support implemented while the cutting wire passes, the cutting step comprising successive phases, and each phase comprising:
- a first period for which the reconciliation is carried out at a first speed V ₁ of reconciliation,
- a second period for which the approximation is carried out at a second speed V ₂ of approximation, the second speed V ₂ of approximation being strictly lower than the first speed V ₁ of approximation and the second speed V ₂ of approximation being such that the cutting of the block is carried out in a ductile mode during the second period.

Such a cutting process has the advantage of improving the cutting by limiting the appearance of defects on the slices thanks to the ductile cutting mode which is implemented several times during the cutting step. The use of the first and second approach speeds in the different phases makes it possible to have a cut in ductile mode during each phase preferably while seeking to reach a desired cutting duration.

The cutting process may also include one or more of the following characteristics:
- During each first period, the cutting of the block is carried out according to a fragile mode;
- the first speed V ₁ of approach, the duration T ₁ of the first period, the second speed V ₂ of approach and the duration T ₂ of the second period are determined so as to satisfy the following condition: the successive phases induce the cutting of the block according to a thickness E in a desired cutting duration D _C , the thickness E being less than or equal to the thickness of the block;
- the condition is expressed by the following equation:
;
- the second approach speed V ₂ is equal to αxV _duct , α being a weighting coefficient less than or equal to 1 and V _duct being a ductile speed corresponding to a maximum speed up to which the cutting of the block is done according to the ductile mode the running speed of the cutting wire;
- the cutting process includes the determination of the values of E, of α, of T ₁ , of T ₂ and of V _duct , and the first approach speed V ₁ is determined by solving the following equation: , β being equal to T ₁ /T ₂ ;
- the first approach speed V ₁ , the second approach speed V ₂ , the duration T ₁ of the first period and the duration T ₂ of the second period are determined so as to maximize, for each of the slices, the surface obtained by cutting of the block according to the ductile mode;
the first approach speed V ₁ is determined so as to be as fast as possible while making it possible to maintain the deflection of the cutting wire, at the level of the web, at a value less than or equal to a predetermined threshold;
- the number of successive phases is determined by taking into account data representative of the behavior of a standard cutting wire when cutting a standard block into slices, the data being chosen from: a monitoring of the evolution of the deflection of the standard cutting wire, monitoring of the first or second derivative of the evolution of the deflection of the standard cutting wire, monitoring of the forces exerted on the standard cutting wire, monitoring of the torque of a motor of driving the standard cutting wire, or monitoring the measurement of the diameter of the standard cutting wire;
- the block being made of silicon, for each second period: the second approach speed V ₂ is between 240 μm/min and 1200 μm/min, and a ratio, established between the second approach speed V ₂ expressed in μm/min and the running speed of the cutting wire, expressed in m/s, is less than or equal to a threshold comprised between 10.43 and 16;
- all of the successive phases correspond to an alternation of the first periods and the second periods so that each variation between the first approach speed V ₁ and the second approach speed V ₂ is instantaneous or is carried out according to a predetermined speed ramp .

The invention also relates to a cutting machine for cutting a block into slices, the cutting machine comprising a support for the block, a sheet formed by at least one cutting wire, the cutting machine being configured to implement :
- a relative rapprochement between the sheet and the support while the cutting wire is scrolling,
- one of the successive phases,
and the cutting machine being configured so that each phase comprises:
- a first period for which the reconciliation is carried out at a first reconciliation speed,
- a second period for which the approach is carried out at a second approach speed, the second approach speed being strictly lower than the first approach speed and the second approach speed being such that the cutting of the block is carried out according to a ductile mode during the second period.

Other advantages and characteristics may emerge from the detailed description that follows.

Brief description of the drawings

The invention will be better understood on reading the following detailed description, given solely by way of non-limiting example and made with reference to the appended drawings listed below.

There illustrates, in a perspective view, a cutting machine used in the context of a cutting process according to one embodiment of the invention, this cutting machine being in the process of cutting a block into slices using a cutting wire.

There is a side view of the machine shown in .

There schematically illustrates part of the bonded abrasive type cutting wire.

There schematically represents, for a phase of the cutting process, the evolution of the type in the form of a crenel of a speed of relative approach between a sheet and a support according to a particular embodiment of the invention, this speed of approach varying between a first closing speed and a second closing speed.

There illustrates the evolution of the height, or level, of the surface of a slice using a profilometer, this slice being obtained according to a cutting recipe called "standard recipe" without implementing the cutting process according to the present invention.

There illustrates the evolution of the height, or level, of the surface of a slice using a profilometer, the slice being obtained according to a so-called "improved recipe" cutting recipe corresponding to an implementation of the method of cutting according to a particular embodiment of the present invention.

There illustrates the evolution of a ductile speed corresponding to the maximum speed of relative approach between the sheet and the support up to which it is possible to ensure cutting of the block then formed of silicon according to a ductile mode according to a cutting wire running speed.

There illustrates the evolution, for different second closing speeds, of a ratio between the first closing speed applied during a first period and the second closing speed applied during a second period as a function of a ratio between the duration of the first period and the duration of the second period.

There illustrates the evolution, for different second speeds of approach, of a ratio R as a function of the first speed of approach.

There schematically illustrates the speed of support movement as a function of time when implementing the standard recipe and the improved recipe.

In these figures, the same references are used to designate the same elements.

detailed description

The invention relates to a method for cutting a block 100 into slices. The cutting of such a block 100 is illustrated by way of example in FIGS. 1 and 2. The cutting method described below proposes to improve the cutting of a block 100 into slices using a layer 1001 by appropriately varying a speed at which a relative rapprochement between the layer 1001 and a support 1002 to which the block 100 is attached is achieved.

For example, a sacrificial material 1010, in particular in the form of a layer, can be used to secure the block 100 to the support 1002 in order to ensure adequate maintenance of the block 100 to the support 1002. The sacrificial material 1010 can be part of the support 1002. The sacrificial material 1010 can make it possible to fix/hang the block 100 to the support 1002, thus making the block 100 integral with the support 1002. The sacrificial material 1010 can be an epoxy resin loaded with alumina particles. The sacrificial 1010 material may be such that the epoxy resin is in a cured state such that the alumina particles are evenly distributed in the epoxy resin. The alumina particles can each have a maximum dimension of between 1 μm and 30 μm, and the quantity of alumina particles can be such that the sacrificial 1010 material can comprise from 20% to 40% alumina by volume. The alumina particles make it easier to cut the sacrificial 1010 material when the 1001 sheet penetrates this sacrificial 1010 material in the sense that they make the sacrificial 1010 material harder than if it only comprised epoxy resin.

The block 100 can more particularly be an ingot or a brick.

In the present description, by “between two values”, it is understood that the limits formed by these two values are included in the corresponding range.

In the present description, a ratio or ratio between a and b, a and b each of which can take a corresponding value, corresponds to a/b. Unless otherwise specified as part of the definition of a corresponding ratio, the unit of measure of a corresponds to the unit of measure of b.

A cutting machine 1000, also called cutting equipment, is illustrated according to a specific example in FIGS. 1 and 2. 1010 sacrificial.

The sheet 1001 is intended to be used to cut the block 100 into slices, in particular by sawing. The layer 1001 makes it possible in particular to form, by penetrating into the block 100, grooves which are dug as the approach of the layer 1001 and the support 1002 until delimiting the slices when the layer 1001 leaves the block 100.

The web 1001 can be formed by at least one cutting wire 1004 . Of course, the cutting of the block 100 by the layer 1001 can be ensured when the cutting wire 1004 or the cutting wires 1004 pass. Thus, the cutting machine 1000 can comprise the cutting wire(s) 1004 .

The web 1001 may comprise, in particular at each instant of the running of the cutting wire 1004, longitudinal portions 1003 of the cutting wire 1004 defining in particular a region called “cutting region”. The longitudinal portions 1003 of the cutting wire 1004 are arranged successively and spaced apart so as to define the desired thickness of the slices cut from the block 100. In particular, the longitudinal portions 1003 of the cutting wire 1004 are parallel to each other and are, preferably arranged in the same plane. In this case, the sheet 1001 is formed by the cutting wire 1004 and more particularly by turns of a winding of this cutting wire 1004. Commonly, such a sheet 1001 is also called in the technical field of the invention “yarn sheet”.

Alternatively, several cutting threads 1004 can be used to form the layer 1001, in particular at each instant of their running. In this case, in the ply 1001 can be formed by longitudinal portions 1003 of the cutting threads 1004. In this case, the web 1001 can comprise as many cutting threads 1004 as there are longitudinal portions 1003 . The cutting wires 1004 can each be in the form of a closed loop driven in scrolling by a motor associated with said cutting wire 1004. Each cutting wire 1004 can be such as the cutting wire 1004 described below.

In the remainder of the description, the sheet 1001 formed by a single cutting wire 1004 is used. Of course, this ply 1001 can be replaced by that defined by several cutting threads 1004 .

In particular, the cutting wire 1004 is abrasive. In this case, the 1004 cutting wire may also be referred to as the 1004 bonded abrasive cutting wire.

By “bonded abrasive cutting wire 1004”, part of which is shown as an example in , it is understood a cutting wire 1004 comprising a support wire 1004a, for example made of steel, also called the central core, and abrasive grains 1004b (also called abrasive particles) such as, for example, diamonds. The abrasive grains 1004b are attached to the support wire 1004a. In fact, the abrasive grains 1004b emerge from the surface of the support wire 1004a and have a hardness greater than that of the material of the block 100 to be cut. Thus, the cutting wire 1004 can be a diamond wire.

In the case where the web 1001 is formed by the cutting wire 1004, this cutting wire 1004 locally forms the turns mentioned above, these turns making it possible to delimit the longitudinal portions 1003 . This is particularly what is shown in Figures 1 and 2 where the cutting wire 1004 is wound:
- partly on a first reel 1005 and partly on a second reel 1006, thus allowing the cutting wire 1004 to go back and forth between the first and second reels 1005, 1006, these back and forth defining the running of the wire 1004 cutting,
- partly around first and second guide members 1007, 1008 (also called cylindrical guides or wire guides) which allow at any time of the scrolling of the cutting wire 1004 to form the turns and therefore thus the sheet 1001. The first and second guide members 1007, 1008 can each comprise grooves 1009. Thus, the winding of the cutting wire 1004 around the first and second guide members 1007, 1008 can be ensured using the grooves 1009. The first member 1007 guide, the second guide member 1008, the first reel 1005 and the second reel 1006 can be motorized to ensure the round trips, and therefore the scrolling, of the cutting wire 1004. Of course, the first guide member 1007, the second guide member 1008, the first coil 1005 and the second coil 1006 belong in particular to the cutting machine 1000.

Preferably, block 100 is a block of silicon. The wafers cut from this block 100 are therefore made of silicon and are preferably intended to be used for photovoltaic applications (for example to manufacture photovoltaic cells from the wafers). For cutting silicon wafers, the typical diameter of the support wire 1004a is preferably less than or equal to 120 μm with a spacing between the consecutive turns of the winding around the first and second guide members 1007, 1008 included between 150 µm and 300 µm. The size of the abrasive grains 1004b, in particular defined by their maximum dimension, can be between 5 μm and 9 μm. such a cutting wire 1004 makes it possible, for example, to obtain silicon wafers approximately 180 μm thick. Of course, the spacing between turns can be adapted, for example, to allow slices with a thickness of up to 80 μm to be obtained.

The block 100 of silicon may have a section, in particular square, of 156 mm by 156 mm, the slices then being cut so as to each have two opposite faces of 156 mm by 156 mm. Other dimensions can of course be envisaged, but it is these dimensions of block 100 which are considered subsequently, in particular in the standard and improved recipes described.

For example, the cutting wire 1004 may have a length of 50 km, it is possible to cut with this cutting wire 1004 50,000 slices from different blocks 100, i.e. an average consumption of cutting wire 1004 of 1 m per slice of 156mm by 156mm. Thus, during the cutting, the cutting wire 1004 performs round trips by scrolling between the first and second reels 1005, 1006. For each round trip, the outward journey may consist of scrolling the cutting wire 1004 on a few hundred meters, with a slight offset between the outward length and the return length, for example for 600 m of 1004 cutting wire unwinding from the first reel 1005 on which the majority of the 1004 cutting wire is initially wound when the 1004 cutting wire is new, only 594 m of 100 cutting wire is rewound on the first 1005 spool.

In general, a reduction in the thickness of the slices obtained by cutting is sought because it has an economic advantage by making it possible to form more slices in the same block 100, if necessary as long as this thickness does not affect the performance of photovoltaic cells formed from these slices.

By “running of the cutting wire 1004”, it is understood that the cutting wire 1004 runs at a running speed, the running speed being considered to be the absolute value of the stabilized running speed of the cutting wire 1004. Indeed, when the cutting wire 1004 scrolls back and forth, there are transient states corresponding to decelerations, then stopping of the cutting wire, then acceleration, which are not taken into account.

The running speed of the cutting wire 1004 corresponds in particular to the running speed of the longitudinal portions 1003 of the sheet 1001.

In particular, the running speed of the cutting wire 1004 is unchanged during a step of cutting the block 100.

It has been mentioned back and forth of the cutting wire 1004 so as to scroll it, however the cutting wire 1004 can scroll in one direction when cutting the block 100 if its length is sufficient.

The block 100 slicing process can be implemented as follows. The block 100 is integral with the support 1002. The cutting process includes the step of cutting (FIGS. 1 and 2) the block 100 by the layer 1001 formed by said at least one cutting wire 1004. This cutting step includes the relative bringing together of the web 1001 and the support 1002, this bringing together being implemented while the cutting wire 1004 runs for example according to the round trips mentioned above. Thus, the web 1001 will locally see the longitudinal portions 1003 scroll to tear material from the block 100 with a view to cutting it into fine slices.

By "relative bringing together between the web 1001 and the support 1002", it is understood that the bringing together is implemented by moving the web 1001 which is then mobile relative to the support 1002 and/or by moving the support 1002 which is then mobile relative to the web 1001 for example using suitable moving means that can use a motor. Preferably, it is the support 1002 which is mobile so that the movement of the support 1002 allows it to approach the web 1001, the position of which preferably remains fixed in the frame of reference of the cutting machine 1000. This approximation makes it possible to pass the block 100 through the web 1001 from which it results the obtaining of the slices cut in the block 100 by the web 1001.

Thereafter, when it is mentioned "the rapprochement", it is of course referred to the relative rapprochement between the sheet 1001 and the support 1002.

The cutting step comprises successive phases, in particular also called alternating phases, for example each of these phases is as illustrated in . Each phase, i.e. each of the successive phases, comprises:
- a first period P ₁ for which the reconciliation is carried out at a first speed V ₁ of reconciliation, in particular this first period P ₁ has a duration T ₁ ,
- a second period P ₂ for which the approach is carried out at a second speed V ₂ of approach strictly lower than the first speed V ₁ of approach, the second speed V ₂ of approach being such that the cutting of the block 100 is carried out according to a ductile mode (we also speak of cutting in ductile mode) during the second period P ₂ , in particular this second period P ₂ has a duration T ₂ .
The first speed V ₁ of approach can be equal, if necessary, to a speed of movement of the support 1002 in the direction of the sheet 1001 (the sheet 1001 then remains in a fixed position), to a speed of movement of the sheet 1001 in direction of the support 1002 (the support 1002 then remains in a fixed position), or at a speed taking into account a speed of movement of the support 1002 in a direction oriented towards the web 1001 and a speed of movement of the web 1001 in a direction oriented towards the support 1002 (the sheet 1001 and the support 1002 then move one in the direction of the other). The second speed V ₂ of approach can be equal, if necessary, to a speed of movement of the support 1002 in the direction of the sheet 1001 (the sheet 1001 then remains in a fixed position), to a speed of movement of the sheet 1001 in direction of the support 1002 (the support 1002 then remains in a fixed position), or at a speed taking into account a speed of movement of the support 1002 in a direction oriented towards the web 1001 and a speed of movement of the web 1001 in a direction oriented towards the support 1002 (the sheet 1001 and the support 1002 then move one in the direction of the other). In , V ₁ and V ₂ correspond to movement speeds of the support 1002.

These successive phases make it possible to define an alternative mode during the cutting of block 100.

The successive phases correspond in particular to repeating cycles.

The cutting of a material, such as for example the silicon of block 100, according to the ductile mode is a form of machining making it possible to obtain surface states of excellent quality, sometimes even better than by polishing. In the context of cutting according to the ductile mode using the cutting wire 1004 comprising diamonds as abrasive grains 1004b, it is necessary that the thickness of material removed by a diamond from the cutting wire 1004 during its ductile passage through the silicon is greater than or equal to the thickness of a portion of the block undergoing hydrostatic pressure in front of the diamond. This principle can of course be applied to materials other than silicon.

Cutting according to the ductile mode and according to a fragile mode are known from the literature as shown for example in the document “Diamond machining of silicon: A review of advances in molecular dynamics simulation” by Saurav Goel et al. published in International Journal of Machine Tools & Manufacture 88 (2015) 131-164.

The ductile mode as a mode of material removal is not conventionally used in the context of sawing/cutting with cutting wire but rather in machining to allow grinding or polishing after cutting. This is in particular due to the fact that the complete cutting of the block 100 into slices according to the ductile mode would take too long for the attainable speeds of scrolling of cutting wire 1004 . Indeed, the document “Analytical prediction of subsurface microcrack damage depth in diamond wire sawing silicon crystal” by Liyuan Wang et al. and published in Material Science in Semiconductor Processing 112 (2020) 105015, describes a critical ratio of the speed of movement of a part to be cut (in µm/min) to the speed of travel of the cutting wire (in m/s) for cut only in ductile mode. Thus, the speed of movement of the part to be cut is to be assimilated to the speed of relative approach between the sheet 1001 and the support 1002. This critical ratio defined in (µm/min)/(m/s) can be between 10 ,43 and 16 for a uniform part movement speed of between 240 µm/min and 1200 µm/min. This implies that to cut a total thickness of material of 168 mm (for example thickness of the block 100 plus part of the sacrificial material 1010) in 140 minutes, it would be necessary to use a cutting wire speed 1004 of at least 120 m /s to remain in ductile mode, such a scrolling speed not currently being industrially achievable. A more typically used value is a running speed of the cutting wire 1004 equal to 30 m/s in the context of cutting a block of silicon, but such a running speed would return, to always remain in ductile mode during cutting, to use a speed of relative approach between the sheet 1001 and the support 1002 equal to 450 μm/min. Thus, it would be possible to completely cut a block of silicon into slices according to ductile mode but at the cost of extending the cutting time to approximately 6 hours instead of the 2 hours typically necessary to cut the block of silicon 156 mm by 156 mm in slices. It is therefore unlikely that the photovoltaic industry will accept such a loss of productivity.

The brittle mode allows rapid tearing of material and causes surface flaking.

It follows that when the block 100 is made of silicon, for each second period P ₂ , the second approach speed V ₂ can be between 240 μm/min and 1200 μm/min, and a ratio, this ratio being established between the second approach speed V ₂ expressed in μm/min and the running speed of the cutting wire 1004 expressed in m/s, is less than or equal to a threshold comprised between 10.43 and 16. This makes it possible to ensure during the second periods P ₂ the cutting of silicon in ductile mode.

In general, the present invention proposes cutting the block 100 into slices according to a cutting recipe which has phases each comprising an alternation between a high speed of relative approach between the sheet 1001 and the support 1002, and a slow speed of approach. relative between the sheet 1001 and the support 1002 while making it possible to achieve a cutting duration D _C desired for all of the phases. The slow speed corresponds to the second speed V ₂ of approach ensuring a cutting of the block 100 according to the ductile mode and the high speed corresponds to the first speed V ₁ of approach which is considered as a high speed compared to the slow speed.

Whereas one might think that a repetition of alternations of first approach speed V ₁ and second approach speed V ₂ could negatively alter the surface appearance of the slices cut due to an inhomogeneous cut in the time, it turns out on the contrary that it tends to improve the overall surface appearance with a favorable effect on the mechanical resistance of the wafers, in particular of silicon, obtained at the end of the cutting of the block 100.

The successive phases make it possible to cut a part of the block 100 according to the ductile mode using the second periods P ₂ and a part of the block 100 more quickly during the first periods P ₁ to ultimately avoid lengthening significantly the cutting time of block 100.

In order to allow rapid cutting during the first periods P ₁ , block 100 can be cut during each first period P ₁ in fragile mode, also called fragile mode cutting.

In particular, for each phase, the transition from one of the first and second periods P ₁ , P ₂ to the other of the first and second periods P ₁ , P ₂ can be instantaneous in the sense that the speed of approach between the tablecloth 1001 and support 1002 instantaneously passes either from the first approach speed V ₁ to the second approach speed V ₂ , or from the second approach speed V ₂ to the first approach speed V ₁ . In this case, depending on the order of the first and second periods P ₁ , P ₂ for the phases, the transition from a “so-called first phase” phase which directly precedes another phase “so-called second phase” can be such that the speed of rapprochement between the sheet 1001 and the support 1002 passes instantaneously either from the first speed V ₁ of rapprochement of the first period P ₁ of the first phase to the second speed V ₂ of rapprochement of the second period P ₂ of the second phase, or from the second speed V ₂ approaching the second period P ₂ of the first phase to the first speed V ₁ approaching the first period P ₁ of the second phase.

Alternatively, each phase may include a transitory period between its first and second periods P₁, P₂. During this transient period, the speed of approach between the sheet 1001 and the support 1002 varies, for example gradually in a linear manner, to pass either from the first speed V₁approach to the second V₂approach, or second speed V₂approach at first speed V₁of rapprochement. Thus, during each phase, the passage of one of the first and second periods P₁, P₂to the other of the first and second periods P₁, P₂may be dependent either on a deceleration ramp for the movement of the support 1002, or on an acceleration ramp for the movement of the support 1002. According to this alternative, any phase "so-called next phase" which directly follows another phase called " previous phase" may include, before its first and second periods P₁, P₂, an initial period. The initial period of the corresponding next phase makes it possible to adapt, for example according to an acceleration ramp or a deceleration ramp which can be linear, the speed of approach between the web 1001 and the support 1002 at the end of the previous phase either at first speed V₁reconciliation of the first period P₁of the next phase, i.e. at the second speed V₂reconciliation of the second period P₂of the next phase. In particular, if the previous phase ends with the first period P₁then the initial phase of the following phase allows the passage of the first speed V₁reconciliation of the first period P₁from the previous phase to the second speed V₂reconciliation of the second period P₂of the next phase, and vice versa if the previous phase ends with the second period P₂then the initial phase of the following phase allows the passage of the second speed V₂reconciliation of the second period P₂from the previous phase to the first speed V₁reconciliation of the first period P₁of the next phase.

In other words, the set of successive phases corresponds to an alternation of the first periods P ₁ and of the second periods P ₂ , and therefore to an alternation of the first approach speeds V ₁ and the second approach speeds V ₂ , so that each variation between the first approach speed V ₁ and the second approach speed V ₂ is instantaneous or is carried out according to a predetermined speed ramp corresponding either to an acceleration ramp or a deceleration ramp. These two solutions make it possible to improve the cutting of the block 100 into slices by tending to improve the overall surface appearance of the slices with a favorable effect on the mechanical strength of the slices.

To demonstrate certain advantages of the cutting process, an example of a recipe called "standard recipe" is defined corresponding to what can be done without using the successive phases described and an example of a recipe called an "improved recipe" implementing the cutting method according to a particular embodiment of the invention. For each of the standard and improved recipes, the block 100 considered is a block of silicon and the cutting wire 1004 is a diamond wire. The improved recipe described below is of course a non-limiting example which can be compared to the standard recipe. There illustrates by way of example and schematically, for a web 1001 which remains in a fixed position, the speed of movement of the support 1002 (which then corresponds to the speed of approach at which the relative approach between the web 1001 and the support 1002 is set implemented) as a function of time during the implementation of the standard recipe and during the implementation of the improved recipe.

For the standard recipe, a main cutting phase is used for which the support 1002 approaches the web 1001 at a constant approach speed of 1220 μm/min while the cutting wire 1004 runs at a running speed of 30 m/s. During this main phase, the cutting thread 1004 goes back and forth. For each return trip, 800 m of cutting wire 1004 are unwound for a rewinding of 794.15 m of cutting wire. The total cutting time is, in the case of this standard recipe, 150 minutes in order to cut a thickness of 156 mm from the block 100 plus cut the sacrificial material 1010 to a depth of 9 mm. The main phase has a duration strictly less than the total duration of the cut due to the presence of speeds of approach between the web 1001 and the support 1002 adapted for the entry of the web 1001 into the block 100 and for the exit from sheet 101 of block 101, also called approach speed profiles which are then different.

As part of the improved recipe, the main phase of the standard recipe is replaced by successive phases, each comprising:
- the first period P ₁ for which the first approach speed V ₁ is 2170 μm/min, the duration T ₁ of the first period P ₁ being 65 seconds, and
- The second period P ₂ for which the second approach speed V ₂ is 330 μm/min, the duration T ₂ the second period P ₂ being 124 seconds and the second period P ₂ being after the first period P ₁ .
The number of phases, for example equal to 13, is preferably such that the total duration to allow the implementation of all these phases of the improved recipe corresponds is equal to the duration of the main phase of the standard recipe. Furthermore, globally on all the phases of the improved recipe, it is possible to define an average speed of relative approach between the web 1001 and the support 1002 corresponding to the approach speed of 1220 μm/min of the main phase of the standard recipe so that in the end the succession of the phases of the improved recipe makes it possible to cut the block 100 preferentially according to a duration equal to the duration of the main phase of the standard recipe. According to this improved recipe, each transition from the first period P ₁ to the second period P ₂ (corresponding in the present case to a transitional period as defined above) can be achieved by decreasing the speed of approach, for example in a linear manner, to make it pass from the first approach speed V ₁ to the second approach speed V ₂ . This reduction in the speed of approach is not instantaneous and is done in particular for a transient duration of 368 seconds in the example of this improved recipe. The transitory duration allowing passage from the first approach speed V ₁ to the second approach speed V ₂ can be dependent on a corresponding deceleration ramp associated with the displacement of the support 102. According to this improved recipe, each passage of the second period P ₂ from one of the phases to the first period P ₁ from another of the phases (corresponding in the present case to an initial period as defined above) can be achieved by increasing the speed of approach, for example in a linear manner, to move it from the second approach speed V ₂ to the first approach speed V ₁ , this increase is not instantaneous and is done in particular for a transient period of 61 seconds in the example of this improved recipe. The transitory duration allowing passage from the second approach speed V ₂ to the first approach speed V ₁ can be dependent on a corresponding acceleration ramp associated with the displacement of the support 102. On reading this paragraph, the person skilled in the art understands that the acceleration and deceleration ramps will be dependent on the characteristics of the cutting machine 1000 and will have an influence on the desired cutting time D _C.

For example, the average speed can be defined, considering that the phases are identical, as being equal to (T ₁ xV ₁ +T ₂ xV ₂ )/(T ₁ +T ₂ ) if the transient periods and the initial periods do not are not implemented, otherwise this average speed takes into account the transient periods and the initial periods.

In the examples of the standard and improved recipes, a cutting fluid at a temperature of 20° C. and at a flow of 6000 l/h is preferably used to lubricate the web 1001 during cutting.

Furthermore, both in the context of the improved recipe and the standard recipe, the cutting may include an initial step called penetration of the sheet 1001 into the block 100 prior to the alternate phases of the improved recipe or prior to the phase main part of the standard recipe, and a final step allowing the sheet 1001 to leave block 100 after the alternate phases or after the main phase. These initial and final steps being conventional, the cutting step comprising the successive phases can also comprise these initial and final steps without being limited to the particular example of the improved recipe. These initial and final steps are not described in detail, they nevertheless make it possible respectively to facilitate the penetration of the layer 1001 into the block 100 and the exit of the layer 1001 from the block 100. The initial and final steps are particularly visible in . For the improved recipe, the initial step is represented, according to the direction of time, before the first level at V1 and the final step is represented, according to the direction of time, after the last level at V1. For the standard recipe, it is possible to visualize starting from the left first of all the initial stage, then a drop in the speed of movement of the support 1002 to tend to obtain a horizontal straight line representing the speed of displacement of the support 1002 during the main phase which coincides in with part of the successive phases, then a decrease in the speed of movement of the support 1002 in several stages which corresponds to the final stage. The initial and final stages for the standard recipe correspond to additional and "secondary" stages of the standard recipe.

Between the standard recipe and the improved recipe, an effect is noted on the undulations of visible surfaces obtained on the cut slices adopting in particular the form of more or less deep streaks. This is in particular due to the fact that the deflection of the cutting wire 1004 has a value which oscillates over time during the implementation of the improved recipe whereas it has a smoother evolution during the implementation of the standard recipe. Visually, this is reflected, when implementing the improved recipe, by obtaining silicon wafers on the surface of which a pattern is observed presenting grooves which are attenuated compared to the grooves observable on silicon wafers obtained by cutting according to the standard recipe. This results in notable differences in the measurements by mechanical profilometry carried out, on the one hand, on a slice obtained with the standard recipe and, on the other hand, on a slice obtained with the improved recipe.

There shows an example of surface profilometry carried out on a slice cut according to the standard recipe and the shows an example of surface profilometry performed on a slice cut according to the improved recipe. For these figures 5 and 6, the abscissa represents the displacement in micrometers of the head of a profilometer at the surface of the slice concerned and the ordinate represents the height, in micrometers, of the head of the profilometer during the displacement of its head relative to a value 0 set arbitrarily by a baseline of the profilometer. The waviness on the surface of a silicon wafer obtained with the improved recipe ( ) is attenuated compared to the waviness on the surface of a silicon wafer obtained with the standard recipe ( ). The marked peaks visible in for the standard recipe are, in the case of the measurements carried out within the framework of the improved recipe ( ), rather clipped. Thus, thanks to the improved recipe, it is possible to obtain globally flatter surfaces than in the case of the standard recipe. Obtaining such globally flatter surfaces has the advantage of improving the manufacture of photovoltaic cells from cut silicon wafers: it is thus possible to limit, in comparison with a wafer obtained according to the standard recipe, the appearance defects at the end of the steps of texturization, of deposition of thin layers or of screen printing carried out using a wafer obtained according to the improved recipe. In fact, to obtain good electrical contact between the screen printing lines and the surface of the silicon, it is desirable for this surface to be as flat as possible. Furthermore, the improvement in flatness is an asset in the architectures of tandem photovoltaic cells, multi-junction photovoltaic cells or perovskite photovoltaic cells: the successive-phase cutting process can then be advantageously used in the context of the manufacture of such photovoltaic cells.

For the standard recipe, the ratio between the speed of approach (in µm/min) and the running speed (in m/s) of the cutting wire is 1220/30 or 40 (μm/min)/(m/s) during most of the cut. This value of 40 is much higher than the ductile machining limit defined in the document “Analytical prediction of subsurface microcrack damage depth in diamond wire sawing silicon crystal” by Liyuan Wang et al. and published in Material Science in Semiconductor Processing 112 (2020) 105015, which in this case would be 14.4 (μm/min)/(m/s). This ductile machining limit value can of course vary depending on the equipment, and a weighting coefficient can be applied to it in order to take into account a safety margin. This confirms that when cutting according to the standard recipe, the cutting is carried out mainly or even only in fragile mode, which creates many spalls on the surface of the silicon. With the improved recipe, the ratio between the approach speed (in µm/min) and the running speed (in m/s) of the cutting wire alternates between 2170/30 = 72.3 (μm/min)/(m /s) for the first speed V₁approach and 330/30 = 11 (μm/min)/(m/s) for the second speed V₂of reconciliation. According to the improved recipe, part of the cutting is therefore carried out under ductile machining conditions. By implementing the recipe improved it is possible to obtain silicon wafers with surfaces comprising smooth ridges with little chipping.

Another advantage of cutting according to the improved recipe is that it limits subsurface damage.

Beyond the fact that the surface appearance of the cut slices is generally improved thanks to the cutting step comprising the successive phases, it turns out that the implementation of these successive phases during the cutting step , in particular according to the improved recipe, has a positive effect on the mechanical properties of the wafers obtained in particular according to the improved recipe and compared to the silicon wafers obtained according to the standard recipe. Indeed, for the silicon wafers obtained according to the cutting process comprising the successive phases, these wafers show an improvement in their breaking stress (measured in MPa) in four-line bending, parallel or orthogonal to the streaks left by the diamonds of the cutting wire 1004 during cutting, in comparison with a cutting which would have replaced the successive phases known as alternating approach speed (as in the case of the improved recipe) by a constant approach speed (as in the case of the recipe standard) inducing a cutting of the slices in the block 100 according to the fragile mode.

Thus, the cutting process according to the present invention also makes it possible to improve the mechanical behavior of the slices by making them stronger. It is then even possible to consider cutting thinner slices while maintaining a satisfactory robustness of these slices: from the same block 100 it is possible to cut more slices while maintaining a satisfactory slice breakage rate after cutting, this makes it possible to increase productivity and reduce the production costs of such slices.

Preferably, the first approach speed V ₁ , the duration T ₁ of the first period P ₁ , the second approach speed V ₂ and the duration T ₂ of the second period P ₂ can be determined so as to satisfy the following condition : the successive phases induce the cutting of the block 100 according to a thickness E in a duration D _C of desired cutting. The thickness E is less than or equal to the thickness of the block 100 in particular according to the direction of penetration, also called direction of cutting, of the layer 1001 in the block 100 during the cutting of the block 100. In other words, the phases successive are such that the approximation is implemented, for example according to the average speed of approximation, so as to reach a desired cutting time for a thickness E cut from the corresponding block 100. This has the advantage of allowing cutting according to the ductile mode intermittently taking into account the desired cutting duration D _C . Thus, for example, it is possible to cut according to the improved recipe a thickness of material in the block 100 in a time identical to the cutting of this same thickness during the main phase of the standard recipe. This therefore has the advantage of improving the surface appearance of the cut slices without affecting the cutting rate of the slices in the block 100.

The successive phases, including of course if they are present the transitory periods and the initial periods, can be such that they make it possible to cut the major part of the block 100. This makes it possible to increase in the end the surface obtained by cutting according to the ductile mode for each of the slices. Thus, the thickness E can be greater than or equal to 70% of the thickness of the block 100 (also called “total thickness of the block 100”) and preferably strictly greater than 90% of the thickness of the block 100.

Of course, during the cutting process, the cutting step will ultimately make it possible to cut the block 100 into slices and, if necessary, part of the sacrificial material 1010. In this case, the total thickness of material cut during the implementation of the cutting process can be strictly greater than E and also equal to the sum of the total thickness of the block 100 parallel to the block cutting direction 100 and a predetermined depth of grooves formed by the sheet 1001 in the sacrificial material 1010 after the sheet 1001 has left the block 100 in order to finalize in a clean manner the cutting of the slices which can then be detached from the sacrificial material 1010.

Thus, the successive phases can be such that they make it possible to partially cut the block 100 in ductile mode and in brittle mode. Other parts of the block 100 can be cut according to the fragile mode, for example if necessary when specific approach speed profiles are applied when the sheet 1001 enters the block 100 and when the sheet comes out. 1001 of the block 100. Alternatively, the successive phases can make it possible to cut the block 100 according to its entire thickness, thus separating it into slices.

The condition mentioned above can be expressed by the following equation:
which is referenced eq. 1 thereafter.
Of course, it results from what has been described above that, in this equation eq. 1, T ₁ is the duration of the first period, T ₂ is the duration of the second period, V ₁ is the first speed of approach, V ₂ is the second speed of approach, E is the thickness cut from the block by the implementation of the successive phases and D _C is the desired cutting duration corresponding in the end to N times T, T being the sum of T ₁ and T ₂ , with N representing the number of phases, that is to say the number of alternations between the first approach speed V ₁ and the second approach speed V ₂ . This makes it possible, knowing the thickness E and the desired cutting duration D _C , to find the appropriate parameters, in particular by fixing some of them and/or by calculating some of them from fixed parameter(s) ( s) and/or calculated beforehand, namely T ₁ , V ₁ , T ₂ , V ₂ to obtain a sufficiently fast cutting while improving the surface appearance of the cut slices. Moreover, it results from this equation eq. 1 that:
- the successive phases are such that, for each phase, the transition from the first period P ₁ to the second period P ₂ , or from the second period P ₂ to the first period P ₁ , is instantaneous, and
- each phase which directly follows another phase directly implements the first period P ₁ if this other phase ends with the second period P ₂ or directly implements the second period P ₂ if this other phase ends with the first period _P1 .

In particular, the phases can be such that the first periods P ₁ all have the same duration T ₁ and that the second periods P ₂ all have the same duration T ₂ . In this case, the phases can, within the framework of the realization according to the equation eq. 1, repeat at regular intervals equal to T with T = T ₁ + T ₂ .

As mentioned previously, for each phase of the successive phases, the first period P ₁ can precede the second period P ₂ or the second period P ₂ can precede the first period P ₁ . However, it is judicious that each of the successive phases is such that the first period P ₁ is before the second period P ₂ because the first speed V ₁ of approach is faster than the second speed V ₂ of approach, which makes it possible to avoid, at the beginning of the implementation of these successive phases, a floating of the cutting wire 1004 .

As mentioned above, the thickness E cut in block 100 may be strictly less than the total thickness of material to be cut parallel to the direction of cutting in block 100. This total thickness of material may correspond to the thickness of the block 100 plus a thickness cut in the sacrificial material 1010 fixing the block 100 to the support 1002. The total thickness of material can therefore be cut in a total duration of the cutting step strictly greater than the desired cutting duration D _C in alternate mode.

It was mentioned above that the cutting during the second periods P ₂ was done according to the ductile mode. In particular, for this, the second approach speed V ₂ is equal to αxV _duct , α being a weighting coefficient less than or equal to 1 and V _duct being a ductile speed corresponding to a maximum speed up to which the cutting of the block 100 is done according to the ductile mode at the running speed of the cutting wire 1004 . Indeed, for a running speed of the cutting wire 1004, as long as the second approach speed V ₂ is less than or equal to V _duct then the cutting can be done according to the ductile mode. This is how the choice of α makes it possible to determine the appropriate value of the second approach speed V ₂ .

The ductile speed Vduct, also called ductile machining speed, is a parameter known to those skilled in the art as demonstrated for example by the document "Analytical prediction of subsurface microcrack damage depth in diamond wire sawing silicon crystal" by Liyuan Wang et al. . and published in Material Science in Semiconductor Processing 112 (2020) 105015. The data in this document relates a relationship between a critical speed of approaching a silicon object to be cut with respect to a cutting wire and a speed scrolling of the cutting wire to cut the object. At or below this ratio, the sub-surface damage is considered to be zero, which corresponds to a situation assimilated to a purely ductile cut. Thus, on the basis of these data, it is possible to deduce values of the ductile speed Vduct in mm/min in the range of running speeds of the cutting wire 1004 usually encountered, that is to say in particular between 10 m/s and 40 m/s. In particular, this deduction can be carried out using a linear adjustment whose equation and regression coefficient are described in more detail in connection with the . The table shown below and referenced table 1 below, gives the ductile speed values Vduct as a function of the running speed of the cutting wire 1004 to cut the block 100 of silicon.

Cutting wire running speed (m/s) Ductile velocity V _duct (mm/min) 10 0.193 15 0.252 18 0.288 20 0.312 25 0.371 30 0.431 35 0.490 40 0.550

There represents in connection with Table 1 the evolution of the ductile speed Vduct (in mm/min) as a function of the running speed of the cutting wire 1004 (in m/s). The ductile velocity Vduct evolving in a linear or almost linear manner, it is possible to establish a regression equation defined by y = 0.0119x + 0.0737 with a coefficient of determination (R2) equal to 0.9995. This regression equation is determined from three points which correspond to data called “Known data” from the document “Analytical prediction of subsurface microcrack damage depth in diamond wire sawing silicon crystal” by Liyuan Wang et al. and published in Material Science in Semiconductor Processing 112 (2020) 105015. This regression equation then makes it possible, by taking x equal to the running speed of the cutting wire 1004 which is an operating parameter of the cutting machine 1000, to determine y corresponding to the ductile speed Vduct to be associated with the running speed of the cutting wire 1004 in the context of the cutting of the block 100 of silicon. The dotted line of the represents solutions of the regression equation.

Thus, by knowing the running speed of the cutting wire 1004, for example defined by the type of cutting machine 1000 used, it is possible to deduce V _duct therefrom, then the choice of the parameter α then makes it possible to set the second speed V ₂ of approximation which can be used in the equation eq. 1.

The equation eq. 1 can also be written as follows:
.

By taking into account the relation linking the second speed V₂reconciliation at the ductile velocity V_ductwe have :

and defining a ratio β = T₁/T₂we have T₁= (β /(1+ β))xT or T₂= (1/(1+ β))xT, which makes it possible to write the ratio between the first speed V₁approach and second speed V₂reconciliation as follows:

which we note eq. 2.

This equation eq. 2 can be written, taking V ₂ = αxV _duct , .
The cutting process can then comprise the determination of the values of E, of α, of T ₁ , of T ₂ and of V _duct for example by choice or by calculation (V _duct being known because it depends on the running speed of the wire 1004 cutting), from which it then follows that the first approach speed V ₁ is determined by solving the equation . This makes it possible to easily determine V ₁ with a view to carrying out the cutting of the thickness E according to the duration D _C of the desired cutting.

It is noted that the equation eq. 2 does not depend on the number of alternations between V1 and V2, that is to say on the number of successive phases. There presents the evolution of the V1/V2 ratio as a function of β for different values of α, at an average approach speed of 1.25 mm/min (which makes it possible to cut 150 mm in 120 minutes) and at a running speed wire 1004 cutting equal to 30 m/s allowing to deduce from table 1 that Vduct = 0.431 mm/min. On this , curve C1 is established by taking α = 0.7, curve C2 is established by taking α = 0.8, curve C3 is established by taking α = 0.9, curve C4 is established by taking α = 1 Thus, considering α = 0.8 and β = 1 (i.e. T1 = T2 = T/2), the V1/V2 ratio is approximately 6.26, i.e. V2 = 0.345 mm/min and V1 = 2.16 mm /min. For α = β = 1, the V1/V2 ratio is 4.80, i.e. V2 = 0.431 mm/min and V1 = 2.07 mm/min. Thus, it is possible from the equation eq. 2, knowing the running speed of the cutting wire 1004 and the desired average speed of approach which makes it possible to determine the desired DC cutting duration, to determine all the other parameters to be applied to implement the successive phases of the step cutting.

It follows from what has been described above that by increasing the surface of the slices cut according to the ductile mode, this also increases the mechanical strength of the slices (resistance to mono-axial bending) since the SSD of the cut slices is reduced. In this sense, to meet a need to improve the mechanical strength of the cut wafers, it is necessary to maximize the surface of each cut wafer according to the ductile mode. In other words, the first approach speed V ₁ , the second approach speed V ₂ , the duration T ₁ of the first period P ₁ and the duration T ₂ of the second period P ₂ can be preferentially determined so as to maximize, for each of the slices, the surface of the corresponding slice obtained by cutting the block 100 according to the ductile mode, of course while taking into account the thickness E to be cut and the duration D _C of the desired cut. This maximized surface corresponds in particular to a sum of surface portions obtained during the second periods P ₂ .

Intuitively, to obtain a cut of thickness E at the end of the desired cutting duration D _C , the value of V ₂ closest to, or even equal to, V _duct is chosen while seeking to have the value of T ₂ as long as possible, this making it possible to tend to maximize V ₁ while minimizing T ₁ .

More particularly, starting from the equation eq. 2, it is possible to determine, for the cutting implemented by the successive phases of the cutting step, the ratio R between the displacement of the support 1002 (or of the sheet 1001) at the second speed V ₂ of approach and the displacement of the support 1002 (or of the sheet 1001) at the first approach speed V ₁ in the following way: ,
this ratio R does not depend on the number of successive phases and therefore of alternations between the first approach speed V ₁ and the second approach speed V ₂ . There illustrates the evolution of R as a function of V ₁ measured in mm/min for different weighting coefficients α, at an average speed of movement of the support 1002 equal to 1.25 mm/min (i.e. a thickness E of 150 mm cut in a cutting duration D _C of 120 minutes) while the web 1001 remains in a fixed position and at a running speed of the cutting wire 1004 equal to 30 m/s (ie V _duct =0.431 mm/min for silicon). On the , curve C5 is established is taking α = 0.7, curve C6 is established is taking α = 0.8, curve C7 is established is taking α = 0.9, curve C8 is established is taking α = 1 .

This shows a growth of R with the growth of V1. Moreover, this ratio R is maximum when the weighting coefficient α is equal to 1. This confirms that V2 must be determined/chosen so as to remain close to, or even equal to, the ductile velocity Vduct.

In practice, in order to limit the deflection of the cutting wire 1004 and the associated risk of rupture of the cutting wire 1004, V ₁ can be chosen so that the deflection remains less than or equal to a value of, for example, 9 mm or 10 mm . According to a particular example, V ₁ should not exceed 2 mm/min to 2.5 mm/min (this may depend on the cutting machine 1000 used, the mechanical strength of the support wire 1004a, the material cut) which corresponds to speed ratios V ₁ /V ₂ of between 4.6 and 5.8 for α = 1 (i.e. V ₂ = V _duct = 0.431 mm/min for a running speed of the cutting wire 1004 equal to 30 m /s). According to this particular example, the ratio R would then be close to 0.20 to 0.26, i.e. a proportion of the surface of each slice obtained by cutting according to the ductile mode not exceeding 20% to 26% of the total surface obtained from the slice obtained by cutting block 100.

There is therefore a need to have a reliable parameter to be used in determining the first approach speed V ₁ in order to preserve the integrity of the cutting wire 1004 in order to avoid its breaking. This parameter can be a predetermined threshold corresponding to a maximum value of the deflection of the cutting wire 1004, for example between 9 mm and 10 mm. Thus, the first approach speed V ₁ can be determined so as to be as fast as possible while allowing the deflection of the cutting wire 1004 to be maintained, at the level of the sheet 1001 (that is to say in the cutting region ), at a value less than or equal to the predetermined threshold, for example between 9 mm and 10 mm.

The cutting process can be such that the number of successive phases, and therefore the number of alternations between the first approach speed V ₁ and the second approach speed V ₂ , is predetermined. This number of phases may be chosen to tend towards a uniform distribution of ductilely cut areas on the surface of the wafers and/or to help protect the cutting wire 1004 from breaking. Typically, to obtain the uniform distribution, it is sought to implement as many successive phases as possible at alternating speed of approach. Furthermore, considerations related to the dynamics of evolution of the deflection of the cutting wire 1004 under the effect of variations in the speeds of approach (alternations) can lead to restricting this increase, as can practical limitations in the definition of the recipe corresponding on the 1000 machine. We must therefore find a balance. Thus, the number of successive phases can be determined by taking into account data representative of the behavior of a standard cutting wire during cutting, in particular carried out according to the standard recipe, of a standard block into slices. These data can be chosen from: monitoring of the evolution of the deflection of the standard cutting wire, monitoring of the first or second derivative of the evolution of the deflection of the standard cutting wire, monitoring of the forces exerted on the standard cutting wire, a follow-up of the torque of a motor driving the standard cutting wire, or a follow-up of the measurement of the diameter of the standard cutting wire 1004 in-situ. These data can be measured by sensors of the cutting machine 1000 and can be representative according to their value of the suffering of the standard cutting wire, they are therefore particularly suitable for determining the number of successive phases to be implemented in part of the cutting process.

For example, by measuring using inductive sensors the deflection of the standard cutting wire in different positions distributed over the width of the web 1001, orthogonal to the winding direction of the standard cutting wire around the first and second members 1007 , 1008 for guiding to form the layer 1001, data to be processed are obtained. The processing of data in real time makes it possible to smooth the measurements which, in their raw state, present fluctuations linked to the round trips of the standard cutting wire, and to determine the first time derivative of this arrow in order to anticipate possible deviations during cutting and to establish the appropriate number of successive phases. The advantage of using a first time derivative is that its evolution is not subject to the position of the sensor with which the deflection of the standard cutting wire is measured: whatever the position of the sensor, the curve of the first derivative is identical.

Thus, the choice of the number of successive phases can be dictated by a limitation of the amplitude of the variation of the deflection of the standard cutting wire and/or by a limitation of the amplitude of variation of the time derivative of the variation of the deflection of the standard cutting wire, and if necessary by the programming limits of the cutting machine 1000, for example in the number of alternations between V ₁ and V ₂ that it is possible to program on the cutting machine 1000.

Indeed, given the periodic nature of the variations of the first and second approach speeds, the above equations and values do not depend on the number of successive phases implemented during cutting. However, it is clear that an increase in the number of successive phases favors a uniform distribution of the zones obtained by ductile cuts on the surface of the slices. In addition, considerations related to the dynamics of evolution of the deflection of the cutting wire 1004 under the effect of variations in the speed of relative approach between the sheet 1001 and the support 1002 can lead to restricting this increase in the number of phases. successive, as well as practical limitations of definition of the recipe on the cutting machine 1000 considered. The number of successive phases can be between 10 and 20.

Thus, the present invention makes it possible, for example, to produce silicon wafers having an improved surface appearance, with less surface damage, and thus to improve the mechanical properties of the wafers obtained by cutting a block of silicon without increasing the cutting/sawing time.

Although in the present description it has been described a block 100 preferably of silicon, the cutting process is not limited to silicon as material of the block 100. Indeed, although an application referred to above relates to the cutting of the block 100 in silicon slices, the present invention also applies to cut more generally a block of semiconductor material in slices such as mentioned silicon or germanium. The cutting process can also be used to cut a sapphire block into slices or a ruby block into slices. In this case, the corresponding values of the ductile speed V _duct will be determined as a function of the running speed of the cutting wire 1004. Depending on the material of block 100, the cut wafers can be used to form substrates for use in the fields of photovoltaics, microelectronics, optics, and optoelectronics.

It follows from all that has been described above that the invention also relates to the cutting machine 1000 for cutting the block 100 into slices, the cutting machine 1000 comprising the support 1002 for the block 100, the web 1001 formed by said at least one cutting wire 1004, the cutting machine 1000 being configured to implement the cutting step, that is to say to implement: the relative bringing together of the web 1001 and the support 1002 while the cutting wire 1004 scrolls, and the successive phases as described. Of course, the cutting machine 1000 is also configured so that each phase comprises the first period P ₁ and the second period P ₂ . This cutting machine 1000 naturally has the advantage of forming slices while limiting the appearance of defects on the slices.

More particularly, the cutting machine 1000 comprises a control program for the implementation of the cutting process in particular when the program is executed by the cutting machine.

Everything that applies to the cutting process can of course apply to the 1000 cutting machine which makes it possible to implement it, and vice versa.

The cutting process and the machine described clearly find an industrial application in the manufacture of wafers by cutting, for example with a view to manufacturing photovoltaic cells from these wafers.

Claims (12)

Process for cutting a block (100) into slices, the block (100) being integral with a support (1002), the cutting process comprising a step of cutting the block (100) by a layer (1001) formed by at least one cutting wire (1004), the cutting step comprising a relative rapprochement between the web (1001) and the support (1002) implemented while the cutting wire (1004) is scrolling, the cutting step comprising successive phases, characterized in that each phase comprises:
- a first period (P ₁ ) for which the reconciliation is carried out at a first speed V ₁ of reconciliation,
- a second period (P ₂ ) for which the approach is carried out at a second approach speed V ₂ , the second approach speed V ₂ being strictly lower than the first approach speed V ₁ and the second approach speed V ₂ being such that the block (100) is cut in a ductile mode during the second period (P ₂ ). Cutting process according to Claim 1, characterized in that, during each first period (P ₁ ), the cutting of the block (100) is carried out according to a fragile mode. Cutting process according to any one of Claims 1 to 2, characterized in that the first approach speed V ₁ , the duration T ₁ of the first period (P ₁ ), the second approach speed V ₂ and the duration T ₂ of the second period (P ₂ ) are determined so as to satisfy the following condition: the successive phases induce the cutting of the block (100) according to a thickness E in a desired cutting duration D _C , the thickness E being less than or equal to the thickness of the block (100). Cutting process according to Claim 3, characterized in that the condition is expressed by the following equation:
. Cutting process according to any one of Claims 1 to 4, characterized in that the second approach speed V ₂ is equal to αxV _duct , α being a weighting coefficient less than or equal to 1 and V _duct being a corresponding ductile speed a maximum speed up to which the block (100) is cut according to the ductile mode at the running speed of the cutting wire (1004). Cutting process according to Claims 3 and 5, characterized in that it includes the determination of the values of E, of α, of T ₁ , of T ₂ and of V _duct , and in that the first approach speed V ₁ is determined by solving the following equation:
, β being equal to T ₁ /T ₂ . Cutting process according to any one of Claims 1 to 6, characterized in that the first approach speed V ₁ , the second approach speed V ₂ , the duration T ₁ of the first period (P ₁ ) and the duration T ₂ of the second period (P ₂ ) are determined so as to maximize, for each of the slices, the surface obtained by cutting the block (100) according to the ductile mode. Cutting method according to any one of Claims 1 to 7, characterized in that the first approach speed V ₁ is determined so as to be as fast as possible while allowing the deflection of the cutting wire (1004) to be maintained, at the web (1001), at a value less than or equal to a predetermined threshold. Cutting process according to any one of Claims 1 to 8, characterized in that the number of successive phases is determined by taking account of data representative of the behavior of a standard cutting wire when cutting a block standard in slices, the data being chosen from: monitoring of the evolution of the deflection of the standard cutting wire, monitoring of the first or second derivative of the evolution of the deflection of the standard cutting wire, monitoring of forces exerted on the standard cutting wire, monitoring the torque of a motor driving the standard cutting wire, or monitoring the measurement of the diameter of the standard cutting wire. Cutting process according to any one of Claims 1 to 9, characterized in that the block (100) being made of silicon, for each second period (P ₂ ):
- the second approach speed V ₂ is between 240 μm/min and 1200 μm/min, and
- a ratio, established between the second approach speed V ₂ expressed in µm/min and the running speed of the cutting wire (1004) expressed in m/s, is less than or equal to a threshold between 10.43 and 16 . Cutting process according to any one of Claims 1 to 3, characterized in that the set of successive phases corresponds to an alternation of the first periods and of the second periods so that each variation between the first approach speed V ₁ and the second approach speed V ₂ is instantaneous or is carried out according to a predetermined speed ramp. Cutting machine (1000) for cutting a block (100) into slices, the cutting machine (1000) comprising a support (1002) for the block (100), a layer (1001) formed by at least one wire (1004) cutting, the cutting machine (1000) being configured to implement:
- a relative rapprochement between the web (1001) and the support (1002) while the cutting wire (1004) is scrolling,
- successive phases,
characterized in that the cutting machine (1000) is configured so that each phase comprises:
- a first period (P ₁ ) for which the rapprochement is carried out at a first speed (V ₁ ) of rapprochement,
- a second period (P ₂ ) for which the approach is carried out at a second speed (V ₂ ) of approach, the second speed (V ₂ ) of approach being strictly lower than the first speed (V ₁ ) of approach and the second approach speed (V ₂ ) being such that the block (100) is cut in a ductile mode during the second period (P ₂ ).