• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a method of calibrating a furnace for subjecting a sample of semiconductor material to a first annealing for forming the thermal donors, the first annealing successively comprising a rise (P1) in the temperature of the furnace, a first bearing (P2) at a set temperature (Tc) and a descent (P3) in oven temperature. This calibration method comprises the following steps: a) providing a standard part made of the semiconductor material; b) determining the interstitial oxygen concentration of the standard part; c) subjecting the calibration piece to a second thermal donor formation annealing in the furnace, the second annealing comprising furnace temperature rise and fall identical to those of the first annealing and a second plateau at the set temperature (Tc) for a set duration; d) determining the concentration of thermal donors formed in the standard part during the second annealing; e) determining an equivalent annealing time at the set temperature (Tc), corresponding at least to said rise and fall in oven temperature, from the interstitial oxygen concentration, the thermal donor concentration of the standard part and the the set time.
    公开号:FR3045831A1
    申请号:FR1562979
    申请日:2015-12-21
    公开日:2017-06-23
    发明作者:Sebastien Dubois；Jordi Veirman
    申请人:Commissariat a lEnergie Atomique CEA；Commissariat a lEnergie Atomique et aux Energies Alternatives CEA；
    IPC主号:
    专利说明:

    METHOD FOR CALIBRATING A NIGHT OVEN USED TO FORM THERMAL DONORS
    TECHNICAL AREA
    The present invention relates to the manufacture of wafers from an ingot of semiconductor material. More particularly, the invention relates to a method of calibrating an annealing furnace used to create thermal donors in a sample of semiconductor material, such as a piece of ingot or a brick cut in the ingot.
    STATE OF THE ART
    The silicon wafers used for the production of high efficiency photovoltaic cells are generally derived from monocrystalline silicon ingots obtained by the Czochralski (Cz) drawing process. These platelets are preferentially doped n-type, introducing electron donor phosphorus atoms into the silicon. Phosphorus n-doping (P) is preferred to p-type boron doping (B), particularly because of the absence of boron-oxygen complexes which reduce the service life of charge carriers in silicon.
    Phosphorus is incorporated into the silicon before drawing the ingot, when the silicon is in the molten state, by pouring into the molten silicon bath a phosphorus powder or silicon wafers heavily doped with phosphorus. This doping technique has drawbacks, among which is the contamination of silicon. Indeed, the phosphorus powder or the silicon wafers poured into the bath also contain metals and carbon. These impurities are incorporated with silicon at the same time as phosphorus, which leads to contamination of the silicon, first at the bath, then at the ingot. In addition, since the segregation coefficient of phosphorus is low (about 0.35), a significant variation in the phosphorus concentration, and therefore the electrical resistivity, appears on the height of the ingot. However, the performance of photovoltaic cells is dependent on the electrical resistivity. Photovoltaic cells obtained from the ingot will therefore not have the same performance, including the same photovoltaic conversion efficiency. Thus, a portion of the ingot may be unusable, which represents a financial loss for the platelet supplier.
    To avoid these disadvantages, another doping technique involving thermal donors has been developed. It is described in the article ["High Quality Thermal Donor Doped Czochralski Silicon Ingot for Industrial Heterojunction Solar Cells", F. Jay et al., EU PVSEC Proceedings 2015, pp. 316-321]. Thermal donors are agglomerates created from the interstitial oxygen contained in the silicon (ie the oxygen atoms occupy interstitial positions in the crystal lattice), when subjected to a temperature of between 350 ° C and 550 ° C. ° C. Each thermal donor generates two free elections, which causes a variation in the electrical resistivity of the silicon.
    This doping technique makes it possible to obtain resistivity platelets that are almost identical and contain only very few impurities. First, a monocrystalline silicon ingot is crystallized from a molten silicon bath using the Czochralski process. The silicon used to prepare the bath is intrinsic and no dopant has been intentionally introduced into the bath. Under these conditions, the resistivity of the ingot depends only on the concentration of thermal donors. The concentration of interstitial oxygen and the initial concentration of thermal donors, formed during the crystallization, are then measured on the height of the ingot. It can then be calculated, for each height of the ingot, a concentration of additional thermal donors to create to achieve a target resistivity. These additional heat donors are formed during annealing at 450 ° C. For each height of the ingot, the annealing time required to obtain the additional thermal donors is calculated, knowing the concentration of interstitial oxygen. The ingot is then cut into sections. Each section corresponds to an annealing time, since the calculated annealing times are substantially identical in the same section. Finally, each piece of ingot is subjected to annealing at 450 ° C for the corresponding duration, before being cut into platelets. The most critical step of this platelet manufacturing process is the annealing at 450 ° C of the different ingot sections. It is necessary to avoid sudden changes in temperature when a section is introduced into the annealing furnace, then out of the oven, because the section can break. It is therefore necessary to gradually increase the temperature after introducing the section in the oven, and to decrease it just as gradually before extracting the section of the oven. The problem is that it is difficult to know the exact duration of formation of thermal donors. Indeed, thermal donors are also formed during the heating phase (350 ° C to 450 ° C) and the cooling phase (450 ° C to 350 ° C) of the oven. The amount of additional thermal donors formed during the annealing then differs from that calculated from the target resistivity and the target resistivity is ultimately not achieved.
    SUMMARY OF THE INVENTION
    There is therefore a need to provide a method of calibrating a furnace, to precisely control the formation of thermal donors when a sample of semiconductor material is subjected in this oven to a first annealing, the first annealing comprising successively a temperature rise of the oven, a first step at a set temperature and a temperature decrease of the oven.
    According to the invention, there is a tendency to satisfy this need by providing the following steps: a) providing a standard part made of the semiconductor material; b) determining the interstitial oxygen concentration of the standard part; c) subjecting the calibration piece to a second thermal donor formation annealing in the furnace, the second annealing comprising furnace temperature rise and fall identical to those of the first annealing and a second plateau to the set temperature for a duration of depositary; d) determining the concentration of thermal donors formed in the standard part during the second annealing; and e) determining an equivalent annealing time at the set temperature, corresponding at least to said rise and fall in oven temperature, from the interstitial oxygen concentration, the thermal donor concentration of the standard part and the duration deposit.
    In other words, this calibration method makes it possible to determine the annealing time at the set temperature, for example 450 ° C., which corresponds at least partly to the heating and cooling phases of the oven. Knowledge of such information is valuable, particularly in the context of the platelet manufacturing process described above. Indeed, it is thus possible to take into account the thermal donors formed during the heating and cooling phases of the oven, when setting the oven to anneal the ingot sections. The amount of additional heat donors formed by the annealing is then closer to that calculated from the target resistivity. The calibration method therefore makes it possible to make the adjustment of the electrical resistivity by the thermal donors more precise.
    The second temperature stage, between rising and falling temperature of the second annealing, allows more time for the standard part to reach the set temperature. Thus, the rise and fall in temperature of the standard part correspond to those experienced by the sample during the first annealing (the first annealing comprises the first step at the set temperature).
    This second stage also makes it possible to take into account the differences between the temperature of the sample and the set temperature in the determination of the equivalent annealing time, and therefore the possible phenomena of instability of the temperature of the furnace which are passed on. on the sample.
    Advantageously, the standard part has a geometry and dimensions identical to those of the sample. The accuracy of the calibration is thus improved because the standard part and the sample will undergo exactly the same rise and fall in temperature.
    The set duration is advantageously between 30 seconds and 2 hours.
    In a preferred embodiment, the steps a) -e) are implemented for a plurality of standard parts having different values of the set duration, which results in a plurality of values of equivalent duration, and the calibration method further comprises a step of determining the set duration value beyond which the equivalent duration is independent of the set duration.
    The standard part is advantageously cut in an ingot of semiconductor material crystallized according to the Czochralski method, and preferably in a portion of the crystallized ingot last.
    The interstitial oxygen concentration of the standard part can then be determined by measuring the interstitial oxygen concentration on at least one drop resulting from cutting the ingot, said drop being adjacent to the standard part in the ingot.
    The concentration of thermal donors can be determined from a first measurement of the electrical resistivity or the concentration of free charge carriers, performed before the second annealing on a first wafer adjacent to the standard part in the ingot, and of a second measurement of the electrical resistivity or the concentration of free charge carriers, carried out after the second annealing on a second plate taken from the standard part.
    Preferably, the ingot of semiconductor material is obtained from a bath of intrinsic semiconductor material melt.
    BRIEF DESCRIPTION OF THE FIGURES Other features and advantages of the invention will emerge clearly from the description which is given below, by way of indication and in no way limitative, with reference to the appended figures, among which: FIG. the temperature of a furnace and the temperature of a semiconductor sample placed in the furnace during a thermal donor formation anneal; - Figure 2 shows another profile, equivalent to that shown in Figure 1, the temperature of the sample during annealing; FIG. 3 represents steps of a calibration method according to the invention; FIG. 4 illustrates one way of obtaining a standard part for carrying out the calibration method according to the invention; FIGS. 5 and 6 respectively show that the temperatures of the sample and the oven during the annealing (relative to the set temperature) are exceeded and in which way this excess can be fully taken into account in the calibration process. of the invention; FIG. 7 illustrates the concentration of thermal donors formed during the crystallization of a Cz-type silicon ingot, as a function of the solidified fraction of the ingot; FIG. 8A represents the step of measuring the electrical resistivity or the concentration of free charge carriers, which precedes the annealing for forming the thermal donors of the standard part; and FIG. 8B represents the step of measuring the electrical resistivity or the concentration of free charge carriers, which follows the annealing of the formation of the thermal donors of the standard part.
    For the sake of clarity, identical or similar elements are marked with identical reference signs throughout the figures.
    DETAILED DESCRIPTION OF AT LEAST ONE EMBODIMENT
    FIG. 1 shows an example of curves of the temperature Tf inside an oven and the temperature Ts in a sample made of semiconductor material, when this sample is subjected to a thermal donor formation annealing. in the oven. The sample may be a piece of ingot or a brick cut in an ingot. The brick has the overall shape of a rectangular parallelepiped and allows to obtain square or pseudo-square wafers, which will then be used to manufacture photovoltaic cells. The ingot section, of cylindrical shape, is rather intended for the microelectrical industry, where the plates are circular in shape.
    The annealing of the sample comprises a heating phase P1 in which the temperature Tf of the oven increases, from room temperature Tamb (for example 25 ° C) to a set temperature Te. In the example of FIG. 1, the set temperature Te is 450 ° C. The heating phase P1 is followed by step P2 at the set temperature Te for a set duration te. The set time t0 is the annealing time, specified during oven programming, during which the oven is expected to be at set temperature Te (450 ° C). This annealing time is timed from the moment when the temperature Tf of the oven, measured for example by a thermocouple, reaches 450 ° C. When it stops, the oven temperature Tf begins to decrease. Indeed, the temperature step P2 is followed by a cooling phase P3, in which the temperature Tf goes from 450 ° C. to room temperature Tamb.
    As can be seen in FIG. 1, the temperature Ts of the sample placed in the oven is not always equal to the temperature Tf inside the oven. During the heating phases P1 and cooling P3 in particular, the temperature curve of the sample is delayed relative to the temperature curve of the oven, because of the thermal inertia of the sample. This delay is symbolized in FIG. 1 by a duration "td", during which the temperature Tf of the oven is equal to 450 ° C. while the temperature Ts of the sample is less than 450 ° C.
    Thermal donors are formed in the sample when its temperature Ts is between Tmin = 350 ° C and Tmax = 550 ° C. Considering that the temperature Ts of the sample does not exceed 450 ° C. in the example of FIG. 1, the thermal donors are here formed only in the temperature range 350 ° C. to 450 ° C. The quantity of thermal donors created during the annealing can be represented by the area S situated under the temperature curve Ts and above the straight line of equation T (t) = Tmin = 350 ° C. In general, the larger the area S, the greater the quantity of thermal donors created (it increases with the temperature and the annealing time).
    It can be seen that part of these thermal donors is formed during the heating phases P1 and cooling P3 of the furnace. The duration during which the temperature Ts of the sample is between 350 ° C. and 450 ° C., ie the actual formation time of the thermal donors, is greater than the set time tc. There is therefore a difference between the amount of thermal donors formed during the set time tc (specified by the operator when programming the oven) and the amount of thermal donors formed by annealing as a whole.
    In addition, during the delay td (P2 level), the formation kinetics of the thermal donors is not that expected, since the temperature Ts of the sample has not yet reached the set temperature Tc. This delay therefore contributes to creating the difference in quantities of thermal donors between the set time tc and annealing as a whole. However, this contribution is generally small compared to that of the P1 and P3 phases.
    Thus, when setting the oven setting the set time tc P2 plateau at 450 ° C, it should be taken into account that the formation of thermal donors will also take place during the heating phases P1 and cooling P3 of the oven and because the temperature Ts of the sample may be different from the set temperature Tc during a portion of the plateau P2. This allows to control more precisely the amount of thermal donors formed during the annealing.
    The formation kinetics of the thermal donors during the P1 heating and P3 cooling phases is difficult to establish, because the exact profile Ts (t) of the temperature of the sample is unknown. It is therefore not possible to simply determine what part of the thermal donors is formed during phases P1 and P3.
    The calibration method described below makes it possible to circumvent these difficulties. It is assumed that the temperature of the sample during annealing is constant and equal to 450 ° C, then determines the duration of annealing at 450 ° C (only), noted below teq, equivalent on the one hand to phases of rise and fall in temperature and secondly to the (possible) differences between the temperatures Ts and Te during the stage P2.
    Figure 2 shows the profile Ts' (t) of the temperature of the section in such a case. This profile is equivalent to the temperature profile Ts (t), varying between 350 ° C. and 450 ° C. and of duration Λτ, represented in FIG. 1. The term "equivalent" used above to designate the duration teq or the profile of Figure 2 means that the amount of thermal donors formed is the same.
    The temperature profile Ts' (t) is in the form of a slot, with instantaneous rise and fall in temperature and a plateau at 450 ° C. with a total duration that is soon equal to the sum of the set duration tc and the duration equivalent of teq: ttôt ~ tc teq (1)
    The equivalent duration teq takes into account, as indicated above, the formation of thermal donors during the phases of rise and fall in oven temperature and any temperature differences during the set time tc. It varies according to the oven used, the heating and cooling conditions, the geometry and the dimensions of the section. In some rare cases, especially when the delay time td is significant, the equivalent duration teq can take a negative value. In other words, the "real" quantity of thermal donors formed during the annealing is less than that of annealing at 450 ° C. (only) during the set time tc.
    The total time of annealing at 450 ° C. is, for example, the time calculated from a "target" concentration of thermal donors (itself deduced from a target value of resistivity) and the concentration of interstitial oxygen in the reactor. sample, according to the method of obtaining platelets described in the introduction.
    The knowledge of the equivalent duration teq of annealing at 450 ° C then makes it possible to determine (using the relation (1) above) the exact value of the set time tc to be programmed to reach the "target" concentration in thermal donors. .
    FIG. 2 represents steps S1 to S5 of a calibration method of the annealing furnace making it possible to determine the equivalent duration teq. In step S1, there is provided a calibration piece made of the same semiconductor material as the sample, so that the formation kinetics of the thermal donors can be the same. The sample and the standard part are for example crystalline silicon. The monocrystalline silicon obtained by the Czochralski process, called "Cz" silicon, is preferred to multicrystalline silicon resulting from a crucible-directed solidification process, because it has a higher concentration of interstitial oxygen. Thermal donors can then be created in large quantities, which improves the accuracy of the calibration.
    Thus, in one embodiment of the calibration method, the standard part is cut in a silicon ingot Cz, crystallized from a molten silicon bath. Preferably, the silicon charge used to form the molten silicon bath is undoped, that is to say in intrinsic silicon, and no dopant has been added to the bath. The silicon of the bath is therefore intrinsic at the time of drawing the ingot. Thus, the resistivity of the material will be influenced by the thermal donors only, which improves the accuracy of the calibration.
    For a semiconductor material other than those obtained by the Czochralski process, the interstitial oxygen concentration of the semiconductor material is advantageously greater than 1017 cm-3. Since carbon is an element that limits the formation of thermal donors, its concentration in this semiconductor material is advantageously less than 1017 cm-3. Step S2 consists in determining the interstitial oxygen concentration [Oi] of the standard part. The interstitial oxygen concentration [Oi] can be measured directly on the standard part, non-destructively, by an infrared spectroscopy technique commonly called "Whole-rod FTIR". This technique, derived from Fourier transform infrared spectroscopy (FTIR), consists in scanning the standard part with an infrared beam. The absorption of the infrared beam by the standard part makes it possible to determine an average concentration of the interstitial oxygen in the standard part.
    Alternatively, when the standard part has been cut in an ingot, the interstitial oxygen concentration [Oi] can be measured by FTIR on a piece of the ingot which is adjacent to the standard part. This makes it possible to preserve the standard part of the manipulations necessary for the measurement of the concentration [Oi], in particular surface polishing. It may then be assumed that the interstitial oxygen concentration measured on the adjacent piece is equal to the interstitial oxygen concentration [Oi] of the standard piece.
    For example, when the standard part is a piece of ingot, the concentration [Oi] can be measured by FTIR on one of the two plates, or slices, located on either side of the standard part, or on a piece only this plate. Preferably, the wafer (or piece of wafer) has a thickness greater than or equal to 100 μm and its surface is polished before performing the FTIR measurements.
    Figure 4 shows the case where the standard part is a brick 40 cut in the ingot. The interstitial oxygen concentration [Oi] of the standard part is then advantageously measured on any of the falls 41 coming from the cutting of the ingot, and preferably on an inner face 410 corresponding to the cutting plane. Again, the FTIR technique can be used (after polishing the inner side).
    Fig. 4 also shows wafer pieces 42 located in the ingot immediately above, below, and on the sides of the standard piece. As indicated above, the interstitial oxygen concentration [Oi] can also be measured on any of these pieces of wafer 42. The oxygen concentration can also be measured on several pieces of wafer 42, for example two located respectively in each case. above and below the standard part, and the concentration [Oi] of the standard part is then assumed to be equal to the average value of the oxygen concentrations measured on the pieces 42.
    Techniques other than infrared spectroscopy (FTIR, Whole-rod FTIR) can be used to measure the interstitial oxygen concentration [Oi] of the standard part. In particular, it can be made use of the technique described in patent FR2964459. This technique can be applied to a silicon wafer (as described in patent FR2964459), a piece of wafer 42, a cutting drop 41 of the ingot or the standard part 40. In step S3 of FIG. the standard part is subjected to a thermal donor formation annealing comprising a rise and a fall in temperature identical to that of the annealing of the sample (see Fig.1). The furnace heating and cooling conditions are therefore the same between the annealing of the sample and the annealing of the standard part. The temperature of the oven during the annealing of the standard part reaches the set temperature Te, set here at 450 ° C., and then drops back down to room temperature.
    The annealing S3 of the standard part further comprises, like the annealing of the sample, a bearing at the set temperature Te between rise and fall in temperature. The set duration of this step is noted below te. A bearing during the annealing S3 allows the calibration method to take into account any possible differences between the temperature of the sample and the set temperature Te during the set time tc ', including that corresponding to the delay td between the temperature curves of the sample and the oven, because the standard part can thus reach more easily the set temperature Tc. This bearing makes it possible in particular to take into account problems of regulation of the temperature of the furnace at the beginning of the step. The duration teq (of annealing at 450 ° C.) is then equivalent, in terms of the quantity of thermal donors formed, to the contributions of the phases of rise and fall in temperature, and any temperature differences during the duration tc '(including the delay td).
    The setpoint duration tc 'of the step in step S3 is not necessarily equal to the setpoint duration tc of the annealing of the sample. Indeed, the objective of this step is only to bring the temperature of the standard part during the annealing S3 of the temperature Ts of the sample, to refine the maximum determination of the equivalent duration teq.
    Preferably, the standard part has the same geometry and the same dimensions as the sample. The standard part and the sample will thus respond in the same manner to the rise and fall in oven temperature (the thermal inertia is the same between the two parts). This significantly improves the accuracy of the calibration.
    Then, in S4, the concentration of thermal donors [DT] formed during annealing S3 of the standard part is determined. The thermal donors formed during the annealing S3 can be added to the thermal donors initially present in the standard part and which were generated during the cooling of the ingot. The standard piece may also be subjected to a preliminary anneal to destroy these initial heat donors, for example at 650 ° C for 30 min.
    The concentration [DT] is preferably obtained from two values of the electrical resistivity or the concentration of free charge carriers, respectively measured before and after the thermal donor formation annealing S3. Indeed, the appearance of thermal donors causes a variation in the concentration of free charge carriers, and therefore the electrical resistivity. By "free charge carriers" is meant charge carriers (i.e. electrons or holes) that are free to flow in the semiconductor material and are responsible for the electrical conductivity of the material. Their concentration depends on the initial doping level of the semiconductor material (of type p or n), which has been accentuated or compensated by the formation of thermal donors ("electron donor" defects) in concentration [DT],
    A first measurement of the resistivity - or the concentration of free charge carriers - is therefore performed before annealing S3 and a second measurement of the resistivity - or the concentration of free charge carriers - is carried out after annealing S3. The electrical resistivity can be measured by the four-point method, the Van der Pauw method, by inductive coupling or derived from the measurement of the eddy current. The concentration of free charge carriers can be measured by Hall effect, by photoluminescence, by the analysis of the absorption of infrared radiation by the free charge carriers (called FCA analysis, for "Free Carrier Absorption"). ), or deduced from CV measurements.
    The first and second measurements are, in one embodiment of step S4, performed in the same region of the surface of the standard part (40). In an alternative embodiment, the first measurement is made on a wafer adjacent to the standard part in the ingot, for example that immediately above or below the standard part (or both and the average value is then calculated), and the second measurement is performed on a wafer taken from the surface of the standard part or inside the standard part. The same wafer can then be used for the measurement of resistivity and the determination of the oxygen concentration [Oi] (step S2).
    When a resistivity measurement is performed outside the standard part, i.e. on an adjacent piece of the ingot, the measured resistivity value can be extrapolated to determine the resistivity of the standard part. It is the same for a measurement of the concentration of free charge carriers outside the standard part.
    The thermal donor concentration [DT] can be calculated from measured values of electrical resistivity (or free charge carrier concentration) using the usual relations of resistivity and charge carrier mobility. However, it is preferable, as indicated in patent FR2964459, to modify the relationship of mobility to take into account the influence of thermal donors.
    Finally, in step S5, the total time of the annealing at Te = 450 ° C, equivalent in terms of activation of thermal donors annealing implemented in step S3 (rise and fall in oven temperature, phases temperature instability, plateau at 450 ° C), is determined from the interstitial oxygen concentration [Oi] and the concentration of thermal donors [DT] of the standard part, respectively determined in steps S2 and S4.
    The duration ttot can be calculated using a relation drawn from the article ["Kinetic formation of oxygen thermal donors in Silicon", Wijaranakula C.A. et al., Appl. Phys. Lett. 59 (13), pp. 1608, 1991]. This article describes the kinetics of formation of thermal donors in silicon by annealing at 450 ° C. According to the aforementioned article, the concentration of thermal donors [DT], the initial concentration of interstitial oxygen [Oi] and the annealing time at 450 ° C are linked by the following relation:
    (2) with C the diffusion coefficient of the interstitial oxygen at 450 ° C (C = 3.5 × 10-19 cm 2 / s).
    By replacing the terms [DT] and [Oi] of relation (2) above by the values measured in steps S2 and S4, respectively, a value of t equal to the total duration is obtained.
    To calculate the total time ttot, the relation (2) above is preferred because a set temperature Te equal to 450 ° C is preferably chosen to anneal the ingot samples, in order to adjust their resistivity. The temperature of 450 ° C is indeed the one at which the formation kinetics of thermal donors is best controlled. In addition, it constitutes a good compromise between the formation speed of the thermal donors and the maximum concentration obtained.
    Alternatively, the duration ttot can be determined using other mathematical expressions or charts, giving the concentration of thermal donors [DT] as a function of the annealing time t at 450 ° C, for different values of the concentration. in oxygen [Oi].
    For a different setpoint temperature of 450 ^ 0, the mathematical expressions and the charts can be adapted in particular thanks to the teachings of the article ["Effect of oxygen concentration on the kinetics of thermal donor formation in Silicon at temperatures between 350 and 500 ° C, Londos CA et al., Appl. Phys. Lett. 62 (13), pp. 1525, 1993]. This article also describes the kinetics of formation of thermal donors in silicon, but for annealing temperatures of between 350 ° C. and 500 ° C.
    The equivalent duration teq of annealing at 450 ° C. is then determined from the total duration ttot and the set duration tc ', subtracting the set duration tc' from the total duration ttot (see relation (1)). Alternatively, the determination of the equivalent duration teq can be carried out in a single calculation step, from the oxygen concentration [Oi], the concentration of thermal donors [DT] and the set time tc.
    The steps S1-S5 of the calibration method are not necessarily implemented in the order that has just been described with reference to FIG. 3. In particular, step S2 for measuring the oxygen concentration [Oi ] can be performed after annealing S3 of the standard part, rather than before. Indeed, the annealing of formation of the thermal donors has practically no influence on the concentration of interstitial oxygen. The oxygen concentration [Oi] can therefore be measured on a wafer taken from inside the standard part, after the standard part has been annealed S3. Step S2 could also be performed after step S4 for determining the concentration of thermal donors formed by annealing S3.
    Although they are infrequent, deviations between the sample temperature and the set temperature Tc may be present at the end of the sample annealing time tc. To take into account these late differences, the set time tc 'of annealing S3 must be at least as long as the duration of the phase concerned by these temperature differences.
    The set time tc 'of the annealing S3 may vary depending on the shape and dimensions of the standard part (in particular if they differ from those of the sample), the interstitial oxygen concentration of the semiconductor material. and the oven used to carry out the annealing. For example, a time tc 'low can be chosen when it is known that the oven is rather stable in temperature. The calibration process then gains in execution time.
    Without particular indications, the set time tc 'chosen for the calibration will advantageously be chosen between 30 seconds and 2 hours. This range constitutes a good compromise between a sufficiently long duration so that the temperature in the volume of the reference piece is stabilized (at the set temperature) and a short duration so that the quantity of thermal donors obtained during the bearing of the piece standard is not preponderant in comparison with that obtained during the phases of rise and fall in temperature. The ratio of the quantity of thermal donors obtained during the step on the quantity of thermal donors obtained during the phases of rise and fall in temperature is advantageously less than 10.
    In certain annealing furnaces and / or for certain sample sizes, the temperature Ts of the sample may not be stable for a long time during the P2 stage, because of problems of regulation of the oven temperature.
    For example, the temperature Ts of the sample may exceed the set temperature Tc at the beginning of stage P2, shortly after the phase P1 of temperature rise. The calibration method makes it possible to take into account at least in part such an overshoot (as any temperature difference with the set temperature Tc during the set time tc '), and completely if the set time tc' of the annealing S3 is greater than or equal to the duration of the overtaking. However, since the duration of the overshoot is not known, it can not be guaranteed that the desired set time tc 'will take into account the overshoot as a whole.
    To remedy this, the calibration method (i.e. steps S1 to S5) can be implemented several times with a plurality of standard parts, preferably between 4 and 10 times. The standard parts used are advantageously identical at all points (material, shape and dimensions). Different values of set time tc 'are chosen for the annealing S3 of the standard parts. Except for the set time, annealing S3 is the same for all standard parts. By way of example, FIG. 5 shows, in step with stage P2, an overflow phase 50 in which the temperature Tf of the oven and the temperature Ts of the sample exceed the set temperature Tc. Six different values of set duration, noted tcT at W, are chosen between 30 s and 2 hours. The values tcT to W are chosen so as to cover a wide range of time, so that at least one of them is greater than the duration of the overrun. 11 is then obtained a plurality of values of equivalent duration teqi-teq6 associated with the different standard parts and the corresponding set times tcT to tC6 '. Then, in a last step, it is determined what is the minimum set duration, beyond which the equivalent duration teq is constant. The determination of the minimum set time can be carried out after having plotted on a graph the equivalent duration values obtained, as a function of the corresponding values of the set duration.
    FIG. 6 is a graph teq (tc ') corresponding to the example of FIG. 5. It can be seen from this graph that the equivalent duration teq is independent of the set duration tc', from tc '= W. Indeed, as seen in Figure 5, "tc4" is the lowest of the six set-point values covering the overflow in its entirety.
    Thus, the determination of the minimum set time guarantees that a temperature overrun (and, more generally, any difference between the set temperature Tc and the temperature of the sample) at the beginning of the plateau is taken into account in its entirety. The minimum set time can be determined for each oven and / or sample size.
    It should be noted that it is preferable to determine the minimum set time with large standard parts, since the problem of instability of the temperature is amplified with such parts. The same minimum set time can then be used for smaller parts.
    As indicated above, the standard part can be obtained from a Cz silicon ingot crystallized from an intrinsic silicon bath. FIG. 7 is a graph representing the concentration of initial thermal donors [DTi], formed during the crystallization of such an ingot, as a function of the solidified fraction fs of the ingot (expressed in% of the total length of the ingot).
    It is noted on the graph that the concentration of initial thermal donors [DTi] ingot decreases rapidly, and then remain at a low level. From a solidified fraction of about 25%, all concentration values [DTi] are less than or equal to 3.1014 cm-3.
    This zone of low concentration [DTi] is considered to be the portion of the last solidified ingot, called the "foot of the ingot". The ingot foot contains fewer initial heat donors than the ingot head, i.e., the portion of the ingot solidified first, because the ingot foot is in thermal contact with the molten silicon for a shorter time than the ingot head.
    When the standard piece is obtained from a silicon ingot Cz, it is advantageously cut in the ingot foot (fs> 25%). It can then be assumed that the concentration of initial thermal donors [DTi] is negligible compared to the concentration of thermal donors [DT] formed during annealing S3. This makes it possible to simplify the calibration process by not using, in step S4 for determining the concentration of thermal donors [DT], more than one value of the resistivity or the concentration of free charge carriers. Indeed, the concentration of free charge carriers is then assumed to be twice the concentration of thermal donors [DT],
    An exemplary implementation of the calibration method will now be described in connection with FIGS. 8A and 8B. This example relates to a sample of silicon Cz shaped brick, with a square section equal to 156x156 mm2 and 100 mm in height. It is desired to subject this sample to an annealing of thermal donors in air, the annealing comprising a bearing at 450 ° C. The annealing furnace used is heated by means of metal resistors. To avoid the risk of breakage, the following cooling procedure is applied: - the power supply of the metal resistors is cut off after the bearing at 450 ° C and the oven door remains closed for 30 min; - after opening the oven door, the brick is left in the oven for two hours; then - the brick is removed from the oven and left at room temperature.
    A standard piece 80 (FIG. 8A) is extracted from a Cz silicon ingot obtained from an intrinsic silicon charge, of electronic quality. The diameter of the ingot is 9 inches. The standard part 80 is sawn in the portion of the last solidified ingot, located at a fraction fs greater than 50%. The standard part 80 has the same shape and dimensions as the sample (156 mm x 156 mm x 100 mm).
    Two wafers 81 and 81 'are also cut in the ingot, respectively above and below the standard part 80. These wafers have square main surfaces (156 × 156 mm 2) and a thickness of about 500 μm.
    The concentration of interstitial oxygen [Oi] is measured (step S2) in the center of one of the main faces of each wafer 81, 81 'using the technique described in patent FR2964459. The concentration [Oi] is 8.32x1017 cm -3 on the upper plate 81 and 8.38x1017 cm-3 on the lower plate 81 '. The electrical resistivity is also measured at the center of platelets 81 and 81 '. It is equal to 38 Ω.οιτι on the upper plate 81 and 40 Ω.οιτι on the lower plate 81 '. It is assumed that the values of the concentration [Oi] and the resistivity in the standard part 80 are equal to the average of the values measured on the two plates 81 and 81 '.
    Then, the standard part 80 undergoes annealing in the same furnace, with a temperature step at 450 ° C. for a set time equal to 30 min (step S3). The same cooling procedure is used. At the end of the annealing, the standard part is sawn into three parts 810, 820 and 830 (FIG. 8B). The portion taken from the center of the standard part 80 is a square wafer 820 with a thickness of approximately 500 μm. The resistivity after annealing is measured in the center of one of the main faces of this wafer 820.
    The concentration of thermal donors [DT] formed in the standard part 80 during the annealing is calculated from the average value of resistivity before annealing and the resistivity value measured after annealing on the wafer 83. In this example 9, 05.1013 cm-3.
    Finally, the teq annealing time at 450 ° C, equivalent to the heating and cooling phases of the oven, as well as any deviation of the temperature of the sample from 450 ° C, is calculated using the relafon (2), knowing the concentration of thermal donors [DT] and the concentration of interstitial oxygen [Oi]. It is worth in this example 12 minutes.
    Many variations and modifications of the calibration process will be apparent to those skilled in the art. In particular, the calibration method may comprise, before the first resistivity measurement (or concentration of free charge carriers), a step in which the standard part is subjected to annealing at a temperature greater than or equal to 600 ° vs. This annealing at 600 ° C or more suppresses the thermal donors formed during the crystallization of the ingot. The initial value of resistivity (before annealing S3) is then higher, which makes it possible to refine the calculation of the concentration of thermal donors formed during annealing S3. An annealing at a temperature greater than or equal to 600 ° C can also be performed at the end of the calibration process to destroy the thermal donors formed during annealing S3. It is advantageously followed by a measurement of resistivity (or concentration of free charge carriers) in order to improve the accuracy of the calculations.
    Finally, although the calibration method has been described in relation to samples and silicon standard parts, it could be applied to other semiconductor materials, for example germanium or silicon-germanium alloy. Germanium is a potential candidate because oxygen-based thermal donors can also be formed by annealing in germanium.
    权利要求:
    Claims (9)
    [1" id="c-fr-0001]
    1. A method of calibrating an oven for submitting a sample of semiconductor material to a first thermal donor formation annealing, the first annealing successively comprising a rise (P1) in oven temperature, a first bearing (P2 ) at a set temperature (Te) and a descent (P3) in oven temperature, the method comprising the following steps: a) providing (S1) a standard part (40, 80) made of the semiconductor material; b) determining (S2) the interstitial oxygen concentration of the standard part; c) subjecting (S3) the standard part to a second thermal donor formation annealing in the furnace, the second annealing comprising furnace temperature rise and fall identical to that of the first annealing and a second stage to the set temperature ( Te) during a set duration (te); d) determining (S4) the concentration of thermal donors formed in the standard part during the second annealing; e) determining (S5) an equivalent duration (teq) of annealing at the set temperature (Te), corresponding at least to said rise and fall in oven temperature, from the interstitial oxygen concentration, the concentration of thermal donors the standard part and the set duration (te).
    [2" id="c-fr-0002]
    2. The method of claim 1, wherein the standard part (40, 80) has a geometry and dimensions identical to those of the sample.
    [3" id="c-fr-0003]
    3. The method of claim 2, wherein the set duration (te) is between 30 seconds and 2 hours.
    [4" id="c-fr-0004]
    4. Method according to one of claims 1 and 2, wherein the steps a) -e) are implemented for a plurality of standard parts having different values of the desired duration (te), which results in a plurality of values of equivalent duration (teq), and further comprising a step of determining the set duration value beyond which the equivalent duration (teq) is independent of the set duration (te).
    [5" id="c-fr-0005]
    5. Method according to any one of claims 1 to 4, wherein the standard part (40, 80) is cut in an ingot of crystallized semiconductor material according to the Czochralski method.
    [6" id="c-fr-0006]
    6. The method of claim 5, wherein the interstitial oxygen concentration of the standard part (40) is determined by measuring the interstitial oxygen concentration on at least one drop (41) resulting from cutting the ingot, said drop (41). ) being adjacent to the standard part (40) in the ingot.
    [7" id="c-fr-0007]
    7. Method according to one of claims 5 and 6, wherein the concentration of thermal donors is determined from a first measurement of the electrical resistivity or the concentration of free charge carriers, carried out before the second annealing (S3 ) on a first wafer (81, 81 ') adjacent to the reference piece (80) in the ingot, and a second measurement of the electrical resistivity or free charge carrier concentration, performed after the second annealing (S3 ) on a second plate (820) taken from the standard part (80).
    [8" id="c-fr-0008]
    The method of any one of claims 5 to 7, wherein the ingot of semiconductor material is obtained from a bath of intrinsic semiconductor material in melt.
    [9" id="c-fr-0009]
    9. The method of claim 8, wherein the standard part (40, 80) is cut in a portion of the crystallized ingot last.
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3184673B1|2019-01-23|Method for calibrating an annealing furnace used to form heat donors
    EP2697632A1|2014-02-19|Method for determining interstitial oxygen concentration
    FR2964459A1|2012-03-09|METHOD FOR MAPPING THE OXYGEN CONCENTRATION
    FR2908125A1|2008-05-09|PROCESS FOR PURIFYING METALLURGICAL SILICON BY DIRECTED SOLIDIFICATION
    FR3009380A1|2015-02-06|PROCESS FOR LOCATING A PLATELET IN ITS INGOT
    WO2014064347A1|2014-05-01|Method for forming a doped silicon ingot of uniform resistivity
    EP3210006B1|2019-05-08|Method for characterizing the interstitial oxygen concentration in a semiconductor ingot
    FR3075379B1|2019-11-22|METHOD FOR VALIDATING THE THERMAL HISTORY OF A SEMICONDUCTOR INGOT
    WO2013150194A1|2013-10-10|Determination of the concentration of interstitial oxygen in a semiconductor sample
    WO2016062796A1|2016-04-28|Method for characterising the interstitial oxygen concentration in a semiconducting ingot
    WO2017102832A1|2017-06-22|Method for adjusting the resistivity of a semi-conductive ingot during the production thereof
    FR3028266A1|2016-05-13|PROCESS FOR MANUFACTURING ANY-TYPE MONOCRYSTALLINE SILICON BILGE WITH CONTROLLED OXYGEN THERMAL CONCENTRATION CONCENTRATION
    FR3059820B1|2019-06-28|TEMPERATURE MEASURING METHOD
    FR3059821B1|2019-07-12|TEMPERATURE MEASURING METHOD
    FR2978600A1|2013-02-01|METHOD AND DEVICE FOR MANUFACTURING LAYER OF SEMICONDUCTOR MATERIAL
    EP3510639B1|2020-08-05|Method for sorting silicon wafers according to the net doping variation
    FR2927910A1|2009-08-28|METHOD OF CRYSTALLOGENIZING ELECTRICALLY CONDUCTIVE MOLDED MATERIAL.
    FR3045073A1|2017-06-16|PROCESS FOR MANUFACTURING A INGOT OF OXYGEN-ENRICHED SEMICONDUCTOR MATERIAL AND CRYSTALLIZATION FURNACE
    同族专利:
    公开号 | 公开日
    JP2017141142A|2017-08-17|
    FR3045831B1|2018-01-26|
    US10240871B2|2019-03-26|
    US20170176105A1|2017-06-22|
    CN106894093A|2017-06-27|
    EP3184673B1|2019-01-23|
    EP3184673A1|2017-06-28|
    TW201732211A|2017-09-16|
    KR20170074192A|2017-06-29|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US20090261299A1|2008-03-21|2009-10-22|Covalent Materials Corporation|Silicon wafer|
    FR2997096A1|2012-10-23|2014-04-25|Commissariat Energie Atomique|PROCESS FOR FORMING A SILICON INGOT OF UNIFORM RESISTIVITY|
    CN1324664C|2002-04-10|2007-07-04|Memc电子材料有限公司|Process for controlling denuded zone depth in an ideal oxygen precipitating silicon wafer|
    FR2858462B1|2003-07-29|2005-12-09|Soitec Silicon On Insulator|PROCESS FOR OBTAINING THIN LAYER OF IMPROVED QUALITY BY CO-IMPLANTATION AND THERMAL RECEIVER|
    FR2964459B1|2010-09-02|2012-09-28|Commissariat Energie Atomique|METHOD FOR MAPPING THE OXYGEN CONCENTRATION|
    FR3009380B1|2013-08-02|2015-07-31|Commissariat Energie Atomique|PROCESS FOR LOCATING A PLATELET IN ITS INGOT|JP6669133B2|2017-06-23|2020-03-18|株式会社Sumco|Method for predicting thermal donor generation behavior of silicon wafer, method for evaluating silicon wafer, and method for manufacturing silicon wafer|
    US10793969B2|2018-06-27|2020-10-06|Globalwafers Co., Ltd.|Sample rod growth and resistivity measurement during single crystal silicon ingot production|
    US10781532B2|2018-06-27|2020-09-22|Globalwafers Co., Ltd.|Methods for determining the resistivity of a polycrystalline silicon melt|
    EP3623801A1|2018-09-14|2020-03-18|Commissariat à l'Energie Atomique et aux Energies Alternatives|Method for determining the thermal donor concentration of a semiconductor sample|
    法律状态:
    2016-11-21| PLFP| Fee payment|Year of fee payment: 2 |
    2017-06-23| PLSC| Search report ready|Effective date: 20170623 |
    2017-11-21| PLFP| Fee payment|Year of fee payment: 3 |
    2019-12-30| PLFP| Fee payment|Year of fee payment: 5 |
    2021-09-10| ST| Notification of lapse|Effective date: 20210806 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1562979A|FR3045831B1|2015-12-21|2015-12-21|METHOD OF CALIBRATING A NIGHT OVEN USED TO FORM THERMAL DONORS|
    FR1562979|2015-12-21|FR1562979A| FR3045831B1|2015-12-21|2015-12-21|METHOD OF CALIBRATING A NIGHT OVEN USED TO FORM THERMAL DONORS|
    JP2016240061A| JP2017141142A|2015-12-21|2016-12-12|Method for calibrating annealing furnace used for formation of heat donor|
    TW105141309A| TW201732211A|2015-12-21|2016-12-14|Method for calibrating an annealing furnace used to form thermal donors|
    EP16204250.1A| EP3184673B1|2015-12-21|2016-12-15|Method for calibrating an annealing furnace used to form heat donors|
    KR1020160173735A| KR20170074192A|2015-12-21|2016-12-19|METHOD FOR CALIBRATING AN Annealing furnace USED TO FORM THERMAL DONORS|
    US15/385,241| US10240871B2|2015-12-21|2016-12-20|Method for calibrating an annealing furnace used to form thermal donors|
    CN201611180760.0A| CN106894093A|2015-12-21|2016-12-20|To for forming the method that the annealing furnace of Thermal donor is calibrated|
    [返回顶部]