RU2086513C1 - Method of making porous blank for multimode fiber optics with graduated refraction index - Google Patents

Abstract

FIELD: glass materials. SUBSTANCE: invention relates to method of making blanks in processes of vapor deposition of ultradispersed glass. At least at the terminal stage of deposition process oxidizing atmosphere is used, for example, at deposition of blank central part. Use of this atmosphere increases intake of alloying addition, decreases axial deviations and total stabilization of deposition process. Invention proposes also method of flow selection through the burner that optimizes the deposition process. EFFECT: improved method of making. 5 cl, 8 tbl, 6 dwg

Description

 The invention relates to optical fiber fibers and, in particular, to an improved method for manufacturing oxide-doped preforms from which such fibers can be obtained.

 Optical light guide fibers are known to be a core with a higher refractive index surrounded by a cladding with a lower refractive index. Depending on the type of fiber and the characteristics required for it, the radial distribution of the refractive index across the surface of the fiber can be simple or complex. For example, single-mode fibers typically have a refractive index distribution profile in the form of a simple step, i.e. a substantially uniform index of refraction in the core and a sharp decrease in this index at the interface between the core and the shell. On the other hand, to obtain multimode fibers with a wide passband, it is necessary to create a profile with an almost parabolic distribution of the refractive index along the radius of the core in order to reduce the dispersion between the modes (see article: R. Olshansky. Distribution in optical fiber optic fibers. Review of Modern Phys, 1979, N 51, N 2, pp 341-367).

 Based on the current state of the issue, optical fiber can be made in different ways. The present invention relates to such methods where a porous glass preform is formed, which then solidifies. More specifically, this invention relates to methods for depositing deposits of ultrafine glass to produce preforms.

The vapor deposition blanks typically consist of silicon dioxide (SiO ₂ ) selectively doped with at least one metal or metalloid oxide (most commonly referred to as “metal oxide” here) to provide the desired refractive index distribution profile . Currently, in industry, germanium dioxide (GeO ₂ ) is preferred as an alloying additive, although these additives can also be other metal oxides such as titanium oxide, tantalum oxide, lanthanum oxide, antimony oxide, aluminum oxide and the like, and in addition, mixtures of these oxides. Since alloying additives of metal oxides are one of the most expensive components of the raw materials used in the manufacture of optical fiber fibers, it is important to effectively introduce such an alloying additive into the workpiece with minimal losses.

 The burners used in such processes for depositing ultrafine glass have many openings or outlet assemblies. Halide substances, burner fuel, and oxygen reacting with the fuel and these halide substances pass through these openings. Depending on the design of the burner and the characteristics of the deposited substances, one or a mixture of reagents can pass through different openings. In addition, inert gases can pass through some of the openings, individually or in admixture with a reagent, which serve as carriers or means for controlling the shape and temperature profile of the burner flame. A typical burner design is shown in US Pat. No. 3,698,936 to Molzan.

 Ideally, the cured blank should have the same length characteristics. However, in practice, it was found that the hardening process leads to such "axial deviations" along the length of the hardened preform, when the fiber obtained from the tip of the preform has properties different from the fiber obtained from the middle part of the preform, and in turn, the properties of the latter differ from the properties fiber made from the end of the workpiece near the holder.

 For many reasons, such differences are simply not needed. For example, these differences lead to a strong change in the finished product. In addition, being large enough, these differences can lead to unacceptable (discarded) material that does not meet quality control standards for these products. From the point of view of these and other problems, one of the main objectives of the present invention is to reduce the difference between the fibers obtained from different parts of the hardened preform.

In the process of axial steam deposition (APO) of ultrafine particles, the germanium located in the central part of the preform consists of glass particles, which are a solid solution of GeO ₂ and SiO ₂ , while most of the preform is in hexagonal crystalline form.

US Pat. No. 4,627,866 relates to the APO process, where fluorine is added during the deposition of ultrafine glass particles. The use of increased partial oxygen pressures to facilitate the fluorine entry into quartz blanks is described. The reason for the increase in the partial pressure of oxygen in these works is a thorough decomposition of the fluorine-containing substance (for example, CCl ₂ F ₂ , CF ₄ , etc.) so that “fluorine is further effectively added” and “sufficiently small glass particles are synthesized” ( see U.S. Patent No. 4,627,866, column 2, lines 34-39). It is important that these links do not contain the idea of supplying oxygen inside the burner through the outermost fuel passage in an amount exceeding the stoichiometrically necessary for complete oxidation of the fuel leaving the burner. In addition, these references do not describe or propose to reduce the number of unnecessary types of metal oxides that occur when creating a preform of porous glass and which can migrate along the length of this preform.

Although the references cite GeCl ₄ as a “gaseous feedstock for glass” in the “synthesis of ultrafine glass particles”, there is no description of any forms such as germanium / oxygen in metal oxide or the ratio between the proportion of oxygen in the gas flows of the burner and resulting forms such as germanium / oxygen. These references relate to the effect of the oxidizing atmosphere on the deposition of the fluorine-containing substance and there is no description in them or proposals for using such an atmosphere to reduce axial deviations in the workpieces by reducing the number of undesirable types of metal oxide that tend to migrate during subsequent reheating. It is only proposed to use such types of atmosphere due to the fact that the formation of germanium oxide requires the presence of fluorine-containing material and this undoubtedly affects the effect of the oxidizing atmosphere.

The purpose of this invention is to improve the process of vapor deposition of deposits when obtaining simple blanks from ultrafine particles of glass doped with metal oxide. More specifically, the objectives of this invention are:
1) reducing the amount of alloying additives in the form of metal oxide used in the formation of preforms of ultrafine glass;
2) reduction of axial deviations in the hardened billets of ultrafine glass and the resulting billets of ultrafine glass and obtained from them optical light fibers;
3) a decrease in the sensitivity of the process of deposition of ultrafine glass to changes in flow in the burner.

To achieve these and other goals, the invention provides a stabilized process of deposition of ultrafine glass particles, where:
1) the efficiency of introducing metal oxide into ultrafine glass preforms has been increased;
2) the tendency for the movement of such oxides during deposition and solidification in both radial and axial directions is reduced.

Stabilization is achieved by adjusting the flow of oxygen and fuel to the burner during the deposition of ultrafine glass. More specifically, these flows are controlled so that, at least at critical stages of the deposition process, for example, when the center of the core is deposited, the amount of oxygen available for reaction with the fuel stoichiometrically exceeds the amount required for complete oxidation of the fuel. More specifically, in order to achieve complete oxidation conditions control the flow (s) of fuel via a burner, and the flow (s) O ₂ (if needed) directly mixed with the fuel plus with the stream (s), oxygen (if needed) inside the outer fuel stream (in collectively referred to herein as “oxygen within the outermost passage of the burner”).

 Keeping these conditions of complete oxidation, the amount of the raw material containing the dopant necessary to obtain the preforms is reduced and, at the same time, the axial deviations in the preforms and the finished fiber are reduced. In addition, the entire process of introducing a dopant is improved in the sense that it becomes less sensitive to changes (disturbances) in the flow of reagents, oxygen and fuel through the burner.

In FIG. 1 shows the top of a burner for depositing ultrafine glass having a chimney, a ring for internal shielding and oxygen transfer, two sets of holes for fuel and pre-mixing with oxygen, and a set of holes for external shielding; in FIG. 2 the temperature dependence of the proportion of Ge in the form of GeCl ₄ , GeO ₂ and GeO; in FIG. 3 the backscattered electron images of a central sample of the tested preform 3, FIG. 3a, the image is shown at a magnification of 40X and it shows the twist of segments 18 21, in FIG. 3b is obtained at a magnification of 1000X, and it shows the twist of segment 20; in FIG. Figures 4 and 5 show the dependences on the normalized radius (r / a) of the oxygen fluxes (Fig. 4) and GeCl ₄ (Fig. 5) obtained by deposition of the workpiece during the APO. The stoichiometric ratio of oxygen and fuel led to oxidation during deposition along the central axis, which is indicated on the squares, and did not lead to oxidation for the points marked with plus signs; in FIG. 6 is a plot of the square of the normalized radius ((r / a) ² ) of the differential mode delay (DMZ) for fibers prepared from piles at the end or in the middle. The "standard" curve represents the results obtained with non-oxidizing streams corresponding to FIG. 4 and 5, i.e. flows marked by pluses, while the “experimental” curve describes the results obtained with oxidizing flows corresponding to the same figures, i.e. streams marked with squares.

 The invention relates to increasing the efficiency of incorporation of alloying metal oxides into preforms obtained from ultrafine glass, and also to reducing the tendency for these alloying additives to migrate in preforms during deposition and solidification. The following discussion is especially directed to the use of germanium dioxide as a dopant and it is clear that this discussion is equally applicable to other metal oxide dopants, now and possibly in the future for the preparation of optical fiber fibers, including without limitation such dopants, such as titanium oxide, tantalum oxide, lanthanum oxide, antimony oxide, alumina and boron oxide.

In addition, the following discussion takes place from the point of view of the ultrafine glass deposit system, in which SiCl ₄ , GeCl ₄ , O ₂ and CH ₄ serve as raw materials, and the burner has the design shown in FIG. 1. Burner 11 includes a central chimney 13 through which a mixture of SiCl ₄ , GeCl ₄ and O ₂ passes, an inner shielding ring 15 around the chimney 13 through which O ₂ passes, two sets of fuel holes 17 through which the mixture passes CH ₄ and O ₂ (the oxygen used to form this mixture is referred to in the text as “oxygen for pre-mixing”), a set of openings for external shielding 19 through which O ₂ passes. If necessary, the inner shielding ring can be replaced with a set of holes, a porous pad or rim.

 The invention is applicable to systems of deposits and ultrafine glass, including, but not limited to the POI systems now used or further developed, which use different raw materials and / or burners of different designs.

 The invention can be used to obtain different types of optical fiber fibers, including single-mode and multimode. In particular, this invention can be used to reduce axial deviations in the manufacture of a single-mode optical fiber by improving the density of mode field diameters and critical wavelengths, as well as improving dispersion control. This invention is of particular importance in connection with broadband multimode fibers (bandwidth greater than or equal to 600 MHz km), since for this type of fiber a tight tolerance must be imposed on the distribution profile of the refractive index.

 In accordance with this invention, the flows through the ultrafine glass depositor are controlled so that at least during the deposition of the most critical parts of the preform, for example, those most susceptible to migration of the dopant, such as the central part (central axis) of the core, which is known to have a deviation in the concentration of the dopant resulted in an oxidizing atmosphere. If necessary, the oxidizing atmosphere can be used during the deposition of non-critical parts, including the entire deposition procedure.

The torch flows used here are considered oxidative when the number of moles of oxygen available for reaction with the fuel exceeds the number of moles needed to completely oxidize the fuel. For example, for the complete oxidation of one mole of methane, two moles of oxygen are required:
CH ₄ + 2O ₂ CO ₂ + 2H ₂ O,
whereas in the case of hydrogen for each mole of fuel only half a mole of oxygen is needed:
H ₂ + 1 / 2O ₂ H ₂ O.

In the present invention, under conditions of oxidation, the oxygen flows are taken to be directly mixed with the fuel, plus those that are inside the fuel stream, i.e. in the outermost fuel channel of the burner and inside this channel. The function of increased fuel flows is to completely burn the fuel to prevent the occurrence of reducing combustion products, and to completely oxidize SiCl ₄ and GeCl ₄ . The oxygen supplied to the outermost fuel channel and into the interior of the burner channel can be used essentially to control the chemical reaction in the burner flame, while the oxygen supplied outside this same external fuel channel is primarily intended to form the burner flame and it essentially does not provide for adjusting the chemical reaction in the flame.

 For example, in the case of the burner shown in FIG. 1, oxygen for pre-mixing, i.e. passing through the ring for internal shielding and through the chimney, is used to establish, or not, an oxidizing atmosphere. If the ratio of the sum of these oxygen flows to the double stream of methane or half the stream of hydrogen when used as fuel is greater than unity, then the burner works under the oxidation condition provided for in this invention. On the other hand, if the above ratio is less than unity, as in the previous state of this issue, then the burner does not provide an oxidizing atmosphere. For people who know the current state of this issue, it will be obvious what flows should be for burners of other designs.

Under operating conditions and at operating temperatures in the process of steam deposition of ultrafine glass deposits, germanium can exist in three forms: GeCl ₄ (halide, feed form), GeO (monoxide form) and GeO ₂ (dioxide form). GeO ₂ is the desired form, while GeCl ₄ and GeO are the forms that lead to a reduced accumulation of germanium oxide during deposition and to increased migration of germanium oxide during deposition and curing. Reduced accumulation of germanium oxide is simply undesirable, as it increases the cost of raw materials. The unequal migration of germanium oxide along the length of the workpiece of ultrafine glass leads to an inhomogeneous distribution profile of the refractive index in the hardened workpiece. Such heterogeneity of the preform, in turn, leads to heterogeneity of the fibers elongated from different parts of this preform. In particular, in the case of broadband multimode fibers, the manifestation of these differences consists in different bandwidths for fibers of different lengths, which is simply undesirable.

The present invention addresses both of these problems by providing such conditions for deposits of ultrafine glass when the germanium in the preform is stabilized in the form of GeO ₂ . Using FIG. 2, the stability of the deposit provided by the present invention can be monitored. This figure is a graph of the proportion of germanium in each of its three states (GeO (gas), GeCl ₄ (gas) or GeO ₂ ) under equilibrium conditions versus temperature for a given oxygen partial pressure. This graph was obtained from the thermodynamic data presented by L.V. Gurvich, Thermodynamic properties of individual substances, Academy of Sciences of the USSR, vol. 2, 1979.

As this graph shows, as the temperature rises, the proportion of germanium in the state of GeO ₂ increases until, at a sufficiently high temperature, a significant amount of GeO (gas) begins to form. Outside this temperature, the proportion of germanium drops steeply with increasing temperature.

Thus, from the point of view of temperature at equilibrium or under almost equilibrium conditions, the preferred stable region where it is desirable to work is the peak over the area of GeO ₂ . In this temperature range, the fraction of germanium remaining in the form of GeO ₂ is rather large, and the changes in this fraction with temperature are relatively small, so moderate temperature changes cannot lead to significant changes in the fraction of germanium in the form of GeO ₂ .

Of course, the process of deposition of ultrafine glass is not carried out under equilibrium conditions, and temperature is not the only variable affecting this process. However, by analogy with figure 2, this invention provides operating conditions where the process is more stable from the point of view of capture and migration of germanium oxide. Instead of the independent variable being temperature, as in FIG. 2, however, different flows through the burner to deposit ultrafine glass, such as methane stream, oxygen stream for pre-mixing, oxygen stream through the chimney, oxygen stream through the inner shielding ring and the flow of all reagents through the chimney (SiCl ₄ + GeCl ₄ ). By analogy with the reasons applicable above to figure 2, these variables are chosen such that:
1) increase the capture of germanium oxide;
2) reduce the migration of germanium oxide;
3) provide working conditions that are relatively inert to changes in various flows, i.e. in which the process occurs in the “pleasant” and most stable region of flow variation, similar to that in FIG. 2, which is located just below the peak of GeO ₂ .

In view of the dynamic and interdependent nature of the process of steam deposition of ultrafine glass deposits, when choosing the conditions for the movement of flows that provide such results, one must consider the whole variety of interrelated phenomena. The basic chemical equation that determines the conversion of GeO ₂ to GeO is the following:
2 GeO ₂ = 2GeO (gas) + O ₂ (gas) (1)
The Gurvich thermodynamic data indicated above show that this reaction is very temperature dependent. For example, with any partial oxygen pressure and GeO ₂ activity taken, as the temperature rises from 1130 ^° C to 1530 ^° C, the equilibrium partial pressure of GeO (gas) over pure GeO ₂ increases by about 9,200 times. Thus, the influence of the flows flowing through the burner on the temperature is one of the interrelated factors, which plays an important role for the accumulation and migration of germanium oxide.

From a phenomenological point of view, changes in the concentration of GeO ₂ during deposition of ultrafine glass on the preforms can result from: 1) changes in the relative amounts of GeO (gas) and GeO ₂ produced in the smoke stream before the ultrafine glass particles are deposited on the preform, or 2) changes in the amount of oxide Germany, which migrates in the form of GeO (gas) from the surface of ultrafine glass particles, after they are deposited on the workpiece. In other words, GeO (gas) can occur in a smoke stream before precipitation of ultrafine glass or during the decomposition of GeO ₂ in the particles of this glass after they are deposited, for example, during reheating during subsequent passes of the burner during the deposition of ultrafine glass POI. In addition, a certain amount of GeO (gas) generated in the smoke stream can condense on colder particles of ultrafine glass, which are not directly affected by the hotter parts of the flame in the form of GeO ₂ .

In addition to such effects, germanium oxide can be present in particles of ultrafine glass, apparently in a different form. Thus, the thermogravimetric analysis of ultrafine glass preforms showed that in the presence of chlorine or carbon monoxide, the proportion of germanium oxide in such preforms is quite mobile. This indicates that at least part of the total precipitated germanium oxide is in the form of pure GeO ₂ , and not combined with silicon oxide. This pure GeO ₂ can exist in the form of a germanium oxide-enriched film covering particles of an ultrafine type of germanium oxide silica. On the other hand, scanning-transmission electron microscopy analysis showed that some ultrafine glass particles are either made of pure silicon oxide or pure germanium oxide.

 All these considerations play a role in determining the efficiency of the accumulation of germanium oxide and its migration. It was found that for a large accumulation of germanium oxide and a decrease in its migration, this highly interconnected system can be optimized by regulating the flow of oxygen through the burner in order to obtain an oxidizing atmosphere.

 More specifically, it has been found that by increasing the amount of oxygen in one or more streams through the chimney, for pre-mixing or internal shielding, it is possible to improve the accumulation efficiency and reduce the migration of germanium oxide, the effects of increasing oxygen fluxes in the chimney and for internal shielding being most noticeable.

 From equation 1 above and Gurvich’s thermodynamic data, it follows that the amount of GeO (gas) formed in the smoke stream depends on the oxidation state and temperature of this stream. The degree of oxidation of the smoke stream is most strongly determined by the flow of oxygen through the chimney, the flow of oxygen for internal shielding and the flow of reagents through the chimney, with the oxygen flows for pre-mixing and methane having a lesser effect. An increase in oxygen flow through the chimney and for internal shielding results in a flow through the chimney having a higher oxidation state, while an increase in the flow of reagents through the chimney reduces the oxidation state.

The temperature distribution inside the flame is most strongly affected by the flows of methane and oxygen for pre-mixing and for internal shielding, and the increase in oxygen flows for pre-mixing and internal shielding leads to cooling of the flame. In particular, increasing the oxygen flux for internal shielding reduces the temperature of the flux by reducing the oxidation of CH ₄ at the edge of this flux.

 As for the formation of GeO (gas) when heated by a burner, enriched with germanium dioxide of the precipitated particles of ultrafine glass, this effect primarily depends on the surface temperature of the workpiece. The flame temperature is most strongly affected by the flows of methane and oxygen for pre-mixing, and the amount of oxygen for pre-mixing relative to methane is either less or higher than stoichiometric, which leads to a decrease in the flame temperature. On the other hand, it is recorded that the highest surface temperature of the preform occurs when the oxygen stream for pre-mixing with respect to methane is either less or higher than stoichiometric, which leads to a decrease in flame temperature. On the other hand, it is recorded that the highest surface temperature of the preform occurs when the oxygen stream for pre-mixing relative to the methane stream is less than stoichiometric.

In view of these considerations, an increase in the flow of oxygen through the chimney, the flow of oxygen for internal shielding, and at least to some extent the flow of oxygen for pre-mixing leads to a decrease in the amount of GeO (gas) generated either in the smoke stream or on the surface of the deposited ultrafine glass particles . The flows of methane and reagents can also play a role, but to a lesser extent. In particular, a decrease in the flow of methane, as well as the flow of all reagents through the chimney, increase the effective degree of oxidation of the smoke stream, since the ratio O ₂ / (2CH ₄ + SiCl ₄ + GeCl ₄ ) increases as the flows of CH ₄ , SiCl ₄ and GeCl ₄ decrease . In turn, an increase in the oxidation state reduces the amount of GeO (gas) formed.

 Example 1. The effects of capture of germanium oxide and its migration with increasing oxygen fluxes through the chimney for internal shielding and for pre-mixing, described above, are confirmed by the following experiments.

Using the POI process, eight ultrafine glass blanks were prepared. For each blank, the following deposition parameters were systematically changed:
1) the flow of all reagents (SiCl ₄ + GeCl ₄ through the chimney;
2) the ratio of the flows of GeCl ₄ // SiCl ₄ ;
3) the flow of oxygen through the chimney;
4) oxygen flow for internal shielding;
5) methane stream;
6) the ratio of oxygen flows for pre-mixing and methane.

 For all workpieces, the oxygen flow for external shielding was fixed at 7.5 L / min (standard liters per minute).

 For the first four blanks, at each change of flows, three conditions were used (-, 0, +), while the second four blanks used five conditions (-, -, 0, +, ++). Each blank contained 32 34 segments to be tested, with each segment consisting of 15 full (forward and backward) passes during deposition. The test segments were deposited on a rotating spindle (seed rod), which had a length of 70 cm and onto which, as a result of 40 complete passages, the central part of the ultrafine silica glass was deposited. The intervals for the flows of all reagents through the chimney (FF) and methane (CH) used for the first 16 test segments of each billet ("inner half") differed from those used for the last 16 18 segments ("outer half"). The specific values of the used flows are shown in table. fourteen.

 In order to observe the interaction between different values of flows at different variables, a combination of flows was measured for different segments of the workpiece. For eight or nine segments of each billet, average values (0) were established for all gas flows. These segments served as control and allowed to check the variability within the workpiece. In addition, since it is known that the accumulation efficiency of germanium oxide varies depending on the radius of the segment, these control segments were used to exclude the influence of the radius from the data on axial deviations.

Как будет обсуждено ниже в примере 2, эти уравнения 2 и 3 были далее использованы для расчета оптимальных потоков через горелку для отложения ультрадисперсного стекла.In particular, to exclude the influence of the diameter (Dia) from the data on the capture of germanium oxide, the following equations were fitted to the data on germanium oxide for the control segments:
GeO ₂ (Dia) a ₁ + a ₂ • Dia + a ₃ • (Dia) ² (2)
Similarly, to exclude the influence of the diameter from the data on axial deviations (A _x DiS (Dia)), the data for the control sample minus for the final sample (see below) were adjusted for the control segments as follows:

As will be discussed below in Example 2, these equations 2 and 3 were further used to calculate the optimal flow through the burner for the deposition of ultrafine glass.

The preforms were dried and hardened in a curing oven with a silica muffle. During drying / hardening, a mixture of He and Cl ₂ passed down the central part of the preform, and a mixture of He and O ₂ flowed from bottom to top from the muffle around the hardened preform. When the pores in the preform were closed, this flow was replaced by a helium flow. During drying / hardening, the temperature was varied in the range from approximately 920 ^° C to the highest value of approximately 1405 ^° C. The preforms used in this experiment were intended for core beads (including part of the casing), which were subsequently stretched and coated on top with ultrafine particles of the casing to form preforms fiber. However, this was done for convenience, since the fibers in question included a dopant of metal oxide only in the core of the finished fiber. Equally, this invention is also applicable to the processes of deposition of ultrafine glass for preforms that are pulled into fibers without stretching into piles and further deposition of the shell.

After hardening, the workpieces were cut off at locations located approximately 10 inches (approximately 25 cm) from the tip of the workpiece ("middle sample") and approximately 2 inches (approximately 5 cm) from the tip of the workpiece ("final sample"). Microprobe measurements of the concentrations of SiO ₂ and GeO _{2 were} carried out in the usual way. To record changes in the total concentration of GeO ₂ / SiO ₂ , a wide focused scanning of the microprobe was carried out with minute variations within each segment. With such measurements associated with the use of re-scanning during the deposition of ultrafine glass POI, svile was observed. Referring to FIG. 3 images in backscattered electrons demonstrate the effect of such structural types.

Experimental data for eight preforms showed that with an increase in the amount of oxygen in the smoke stream for internal shielding and, to some extent, in the oxygen stream for pre-mixing and CH _{4, the} concentration of retained germanium oxide increases. These data also show that with increasing methane flow, the concentration of this germanium oxide decreases.

These data show that with an increase in the amount of oxygen in the smoke stream and for internal shielding, axial changes in the amount of germanium oxide decrease, which increase with an increase in the flow of reagents through the chimney and the CH ₄ stream.

 These data show that changes in fluxes with an increase in the degree of oxidation of the flame and, in particular, the degree of oxidation in the chimney region leads to greater capture of germanium oxide and smaller axial deviations, which is very desirable.

Example 2. The procedure for the selection (optimization) of flows through the burner to achieve one or more goals:
1) increasing the efficiency of accumulation of dopants;
2) reduction of axial deviations and / or
3) increasing the stability of the process using experimental data of the type obtained in example 1.

 Prepared one or more blanks for testing. In the preparation of such blanks, each of the flows through the burner that needed to be optimized was changed in the entire interval of interest. You can change one stream at a time or change a group of threads at the same time. To reduce the number of workpieces requiring verification, which are necessary for optimization, it is desirable to divide these workpieces into segments, and to change the flows between these segments.

 After preparing the test pieces, it is advisable to harden them, and then measure to determine the concentration of the dopant in different segments. If axial deviations are to be reduced, then dopant concentrations are also measured at different locations along the segment axis.

Для системы, в которой нужно оптимизировать три потока (f₁, f₂ и f₃), подгоняют под результаты концентрации легирующей добавки, используя, например, обычный метод наименьших квадратов. Если воспользоваться коэффициентами, полученными в процессе подгонки (например, a₀, a₁, a₂, a₃, a₄, a₅, a₆, a₇, a₈, a₉), то значения для переменных потоков (например, F₁, F₂ и F₃), лежащие в допустимом интервале для этих переменных потоков и обеспечивающие минимальное накопление легирующей добавки, можно легко вычислить. (Допустимый интервал переменных потоков это область потоков, которые действительно можно использовать на практике из-за ограничений, накладываемых горелкой или другими процессами).According to the analytical part of this procedure, the function of variable flows, for example, is a first-order polynomial in the form:

For a system in which it is necessary to optimize three streams (f ₁ , f ₂ and f ₃ ), the concentration of the dopant is adjusted to the results using, for example, the usual least-squares method. If we use the coefficients obtained in the fitting process (for example, a ₀ , a ₁ , a ₂ , a ₃ , a ₄ , a ₅ , a ₆ , a ₇ , a ₈ , a ₉ ), then the values for variable flows (for example, F ₁ , F ₂ and F ₃ ), lying in the acceptable range for these variable flows and providing minimal accumulation of the dopant, can be easily calculated. (The permissible range of variable flows is the area of flows that can actually be used in practice due to the restrictions imposed by the burner or other processes).

If it is also necessary to minimize axial deviations, then a separate function of variable flows is adjusted for the data on axial deviations. If we use the coefficients obtained as a result of this fitting process, then we can then calculate the values of the variable flows (for example, F ₁ , F ₂ and F ₃ ), lying in the allowable range for variable flows and minimizing axial deviations.

Finally, if it is also necessary to optimize the stability of the process and, in particular, the concentration stability of the dopant, the first derivatives of the concentration function of the dopant are calculated, squared and summed as the "sum of squares" of the concentration stability function of the dopant, then the first derivatives of the concentration functions of the dopant are squared and summed as the "sum of squares" of the stability function (for example, (a ₁ + 2 • a ₄ • f ₁ + a ₇ • f ₂ + a ₈ • f ₃ ) ² + (a ₂ + 2 • a ₅ • f ₂ + a ₇ • f ₁ + a ₉ • f ₃ ) ² + (a ₃ + 2 • a ₆ • f ₃ + a ₈ • f ₁ + a ₉ • f ₂ ) ^2. Then, the values of the variable flows (for example, F1 '', F2 '' and F3 '') are calculated that lie in the allowable range of these variables and which minimize this feature.

 In general, the values of variable fluxes that minimize the accumulation of the dopant cannot be similar to those that minimize the deviation, and, likewise, none of these sets of values can be similar to the one that minimizes stability. However, it was found that each of these optimizations includes an increase in oxygen fluxes, i.e. obtaining an oxidizing atmosphere. That is, all values for variable flows, which increase the capture of the dopant, reduce deviations and increase stability, include the production of an oxidizing atmosphere.

 Thus, the values for the variable flows, ultimately used to produce blanks, can be a compromise between different sets of optimal values, and these compromise values were chosen based on when optimization is most important to obtain a specific product. Alternatively, compromise values can be determined mathematically by simultaneously optimizing three functions using suitable weighted functions to compose different values and function blocks. For example, a function of the concentration of a dopant can be added to inverse values from axial deviations and to the sum of the squares of the functions, each function is multiplied by a weight coefficient, the sum of the weight coefficients is equated to one, and this allows us to calculate the values of variable fluxes that are within acceptable intervals for these variables and minimize these combined functions.

The analytical part of the optimization procedure was applied to the experimental data from Example 1 as follows. First, using equations 2 and 3, as well as the results of experimental measurements from example 1, the following parameters were calculated:
Resid (GeO ₂ ) GeO ₂ GeO ₂ (Pia) (4)
and
A _x Dif (GeO ₂ ) Mid (GeO ₂ ) Tip (GeO ₂ ) - A _x Dif (Dia) (5)
where Resid (GeO ₂ ) is the concentration of the remaining germanium oxide in the average sample after excluding the influence of diameter on the concentration of germanium oxide, and A _x Dif (GeO ₂ ) is the difference in the concentrations of germanium oxide in the average and final samples again after excluding the influence of diameter.

Then, the polynomials of the following form were fitted to the data on Resid (GeO ₂ ) and A _x Dif (GeO ₂ ):
Resid (GeO ₂ ) CO + Cl • C / S + C2 • FO + C3 • 10 + C4 • CH + C5 • PMO / CH + C6 • FF + C7 • C / S • FO + C8 • G / S • 10 + C9 • G / S • CH + C10 • G / S • MO / CH + C11 • G / S • FF + C12 • FO • 10 + C13 • FO • CH + C14 • FO • MO / CH + C15 • FO • FF + C16 • 10 • CH + C17 • 10 • MO / CH + C18 • 10 • FF + C19 • CH • MO / CH + C20 • CH • FF + C21 • MO / CH • FF + C22 • (G / S) ² + C23 • (FO) ² + C24 • (10) ² + C25 • (CH) ² + C26 • (MO / CH) ² + C27 • (FF) ² (6)
and
A _x Dif (GeO ₂ ) DO + D1 • G / S + D2 • FO + D3 • 10 + D4 • CH + D5 • PMO / CH + D6 • FF + D7 • G / S • FO + D8 • G / S • 10 + D9 • G / S • CH + D10 • G / S • MO / CH + D11 • G / S • FF + D12 • FO • 10 + D13 • FO • CH + D14 • FO • MO / CH + D15 • FO • FF + D16 • 10 • CH + D17 • 10 • MO / CH + D18 • 10 • FF + D19 • CH • MO / CH + D20 • CH • FF + D21 • MO / CH • FF + D22 • ( G / S) ² + D23 • (FO) ² + D24 • (10) ² + D25 • (CH) ² + D26 • (MO / CH) ² + D27 • (FF) ² (7)
where FO is the flow of oxygen through the chimney; 1 = 0 oxygen flow for internal shielding; CH = methane stream; PMO = oxygen stream for pre-mixing; FF flow of all reagents through the chimney; G / S is the ratio of the GeCl ₄ / SiCl ₄ streams, with all streams taken in standard liters per minute (l / min). These equations include coefficients for linear two-way interaction effects and quadratic effects. For fitting, you can use higher order polynomials or other functions, and in some cases it may be necessary to adjust the experimental data. However, it was found that for the data obtained in example 1, the polynomials of equations 6 and 7 provide an adequate fit (see below).

 The fitting was carried out using three sets of polynomials: one for the outer halves of the blanks 1 4, another for the inner halves of the blanks 5 8 and the third for the outer halves of the blanks 5 8. The coefficients were determined using the least squares fitting procedure, where, to determine the statistical significance of each term, we used "criterion F". In the table. Figure 5 shows a typical set of coefficients for the outer halves of workpieces 5 8. As can be seen from the table. 5, after fitting, only a limited number of members remain. Each of the three fits included members from 0 to 6, i.e. linear terms, and higher order terms, which were found to be statistically significant, varied from fit to fit.

It was found that the quality of the fit of the regression equations to the Resid (GeO ₂ ) data is quite good with R ² values in the range of 0.81 0.93. The quality of the fit of the equations to the data on A _x Dif (GeO _{2 was} not so good (the values of R ² were in the range of 0.60 0.76), but still remained acceptable.

 To determine the values of the flow parameters that minimized the accumulation of germanium oxide, axial deviations, and stability, for all the flows tested in Example 1, a computer search was performed using the regression equations. The search for the maximum accumulation of germanium oxide was carried out using equation 6 and different sets of coefficients calculated from experimental data, as described above.

The search for minimal deviations was carried out using the modified equation 7. In particular, when searching for flows that give the minimum value of the axial difference in the amount of germanium oxide, you need a minimum in Mid (GeO ₂ ) (middle) Tip (GeO ₂ ) (end), and not minimum in a _x Dif (GeO _2), defined by equation 5. Therefore, to obtain the values of Mid (GeO ₂₎ to the regression equations for a _x Dif (GeO ₂₎ is added to the value a _x Dif (Dia). Equation 3, in particular, was used to calculate the A _x Dif (GeO ₂ ) value in the center of each of the halves, and this value was added to the DO coefficient for that half.

The search for minimum stability was carried out using equation 6 and the following function in the form of "sum of squares":
Sum Sq (GeO ₂ ) (dResid (GeO ₂ ) / dFO) ² + (dResid (GeO ₂ ) / d10) ² ) + (dResid (GeO ₂ ) / dCH) ² + (dResid (GeO ₂ ) / dPMO) ² + (dResid (GeO ₂ ) / dFF) ²
where the first derivatives were calculated using equation 6.

This search provided three sets of optimal flow values. These flow values were combined to provide simple and continuous algorithms suitable for adjusting flows to the burner, such as shown in FIG. 1, upon receipt of a workpiece with a parabolic profile of the distribution of the refractive index. In particular, to obtain the values of the coefficients FO, FI, F2, P _tr , P _tot , A ₁ and A ₂ in the following expressions, the optimal flow values were combined:
Variable flow FO + (FI-FO) [(P _sig -1) / (P _tr -1)] ^AI (P _{sig is} less than or equal to P _tr ) (9)
and
Variable flow FI- (F2-FI) [(P _sig -P _tp ) / (P _tot -P _tr )] A2 (P _{sig is} greater than or equal to P _tr ) (10)
where P _sig , R _tr and P _tot is the number of runs during deposition during manufacture of the workpiece from ultrafine glass, respectively, current, transitional and complete.

 Table 6 shows the coefficients obtained as a result of the optimization process. The methane flows caused by this algorithm turned out to be more significant than those proposed after the optimization procedure. These more significant values were chosen upon receipt of the core workpiece blank, which was sufficiently dense and therefore hardly cracked during deposition. Using the coefficients from the table. 6 allowed to successfully prepare and harden the workpiece.

 In the table. 7 shows a set of non-optimized coefficients to produce a blank of the same type. Comparison of these coefficients with those given in table. 6 shows that for an optimized system, large flows through the chimney are used, for internal shielding and for pre-mixing than for a non-optimized system. In addition, the initial reagent flow through the chimney for the optimized system is slightly lower, the final reagent flow through this chimney is approximately the same, and the methane flow at the beginning is slightly lower and the end is slightly higher.

In terms of oxidation state, the optimized system (see table 6) has the ratios O ₂ / 2CH ₄ = 1.34; 1.34 and 0.89, and for P _sig / P _tot = 0; 0.65 and 1.0, that is, throughout and in most of the deposition process, an optimized system provides an oxidizing atmosphere. Non-optimized system, for comparison (see table. 7), has the ratios O ₂ / 2CH ₄ = 0.95; 0.81 and 0.77 at the same points in the process, i.e. such a non-optimized system is non-oxidative throughout the deposition process.

Using the coefficients from table. 6 and 7, as well as equations 6 - 8, for a number of GeCl ₄ / SiCl ₄ ratios, the values of A _x Dif (GeO ₂ ) and Resid (CeO ₂ ) and SumSq (GeO ₂ ) were calculated. The results are presented in Table 8. Shown in In this table, the estimated refinements A _x Dif (GeO ₂ ) Resid (Geo ₂ ) and SumSq (GeO ₂ ) are significant.

 Example 3. The effect of using an oxidizing atmosphere during deposition of the critical axial part of the workpiece is shown.

 The first blank of the core staff was prepared using the non-optimized algorithm from table. 7 and germanium tetrachloride streams shown by plus signs in FIG. 5. For reference, the oxygen flows through the chimney for this algorithm are indicated by plus signs in FIG. 4.

 Fibers were prepared from the final and middle parts of the preform. The differential mode delay (DMZ) was determined for fibers prepared from the limbs of the part and from the middle part. We calculated the difference between the DMZ for these parts and constructed the dependence of this difference on the square of the normalized fiber radius. The results are shown in FIG. 6 in the form of a curve marked as “standard”. As this curve shows, the workpiece obtained using the algorithm from table. 7, had significant axial deviations of the DMZ parameter, i.e. about 1.23 ns / km.

A second preform was obtained using the same algorithm, but with the flows of oxygen and germanium tetrachloride shown by the boxes in FIG. 4 and 5. While the ratio of O ₂ / 2CH ₄ at the beginning of the deposition for the algorithm from table. 7 was 0.95, i.e. under non-oxidizing conditions, when using an increased flow of oxygen through the chimney, as in FIG. 4, this ratio was 1.03, i.e., an increase in the oxygen flow through the chimney led to an oxidizing atmosphere during deposition of the central part of the preform. As in the case of the first preform, fibers were prepared from the end and middle of the second preform. The difference between the DMZ for these parts was calculated, and the results were constructed in FIG. 6 in the form of a curve marked as “experimental”. As this curve shows, a preform prepared using an oxidizing atmosphere during deposition on the center line had noticeably reduced axial deviations, i.e. deviations of 70% lower than deviations for the workpiece prepared using a non-oxidizing atmosphere.

 In addition to this important result, the capture efficiency of germanium oxide in the second preform was 8% higher than in the first preform. In addition, fiber prepared from a second preform possessed excellent physical properties.

 As this example shows, the use of an oxidizing atmosphere leads to a marked improvement in the deposition process, even if the oxidizing atmosphere is used only in the restrictive range of the deposition procedure.

 By reducing, in accordance with this invention, axial deviations, most of the preform for the optical fiber provides a fiber that satisfies and even exceeds the predetermined technical characteristics. For example, in the case of a billet stock for a multimode core required to produce a multimode fiber with a peak delta of approximately 2% for the manufacture of an optical fiber with a bandwidth of more than 600 MHz km, approximately 65% of the 150 km billet stock of the core could be used before implementing the invention . When using this invention, the percentage increases to about 90%. For a multimode fiber with a peak delta of about 1%, a similar high percentage of use could be achievable with the application of this invention, i.e., when harvesting a 200 mm long core bead to produce an optical fiber with a bandwidth more than 1500 MHz km, a similar high percentage of this workpiece could be used. Such a percentage suggests that a typical measurement length is at least 0.5 to 2.0 km.

 To obtain optical fiber preforms with limited axial deviation, for example, internal plasma deposition methods can be used, however, these methods usually produce preforms from which less than about 30 km of optical fiber can be drawn, and these methods are not used to make a core head and further coating.

Claims (5)

 1. A method of manufacturing a porous preform for a multimode optical fiber with a graduated index of refraction, containing a silicon dioxide core doped with germanium dioxide, by vapor deposition from a gas mixture of reactants and oxygen supplied through the central channel of the burner with the fuel supplied through its peripheral channel, characterized in that through the coaxial channels of the burner located before and after the fuel channel, oxygen is supplied in an amount ensuring complete combustion of the fuel.  2. The method according to claim 1, characterized in that the reagents used are silicon dioxide precursors, germanium dioxide precursors, oxygen and fuel.  3. The method according to claim 1, characterized in that most of the multimode optical fiber is made with a passband greater than or equal to 600 MHz / km, which is at least partially a consequence of the distribution profile of the refractive index, adjustable in the method.  4. The method according to claim 1, characterized in that the fuel passes to the surface of the burner through many holes made along one or more circumferences of the peripheral channel.  5. The method according to claim 1, characterized in that it further includes the step of drying, curing and drawing the workpiece and subsequent coating with a precipitate to form a fiber.

Images