process of assembling parts by homogeneous diffusion welding.

Abstract

Homogeneous diffusion welding assembly process. A process comprising the diffusion welding of a stack of parts, at least two of which are in contact and made of the same alloy, the alloy exhibiting: - an absence of abnormal grain growth following a first reference heat treatment carried out under thermal conditions identical to those of the diffusion welding, - a fully recrystallized microstructure with an average equivalent grain size of less than 50 µm following a second reference heat treatment comprising successive heating of the alloy identical to the heating of the rise phase of the diffusion welding, and quenching, and - an average equivalent grain size at the end of the first reference heat treatment greater than or equal to 1.5 times the average equivalent grain size at the end of the second reference heat treatment. No figure for the abbreviated version.

Description

Title of the invention: method for assembling parts by homogeneous diffusion welding. Scope of the invention

[0001] The present invention relates to a method for producing an assembly, by diffusion welding, of parts made of the same material. State of the art

[0002] Diffusion welding consists of bringing parts into contact and forming an assembly of these parts under the combined effect of compression and heating to a diffusion welding temperature lower than the melting temperature of the alloy constituting the parts, generally greater than 0.7*Tf, Tf being the absolute melting temperature expressed in Kelvin.

[0003] The "fusion welding" process comprises, in succession:

[0004] - a "rising" phase comprising heating and applying a compressive force to the parts to be assembled, - a "stall" phase, involving maintaining the force at the welding-diffusion temperature for an appropriate duration, then - a "descent" phase involving cooling and a progressive elimination of the compression force, for example by depressurization, notably down to atmospheric pressure.

[0005] Diffusion welding is said to be homogeneous when the parts are made of the same alloy.

[0006] The pores defined between the parts in contact are progressively eliminated during the welding-diffusion cycle.

[0007] The properties of the assembly obtained after the descent phase depend in particular on the condition of the surfaces of the parts which are intended to be brought into contact during welding, as well as on the composition and microstructure of the alloy.

[0008] For example, to join two plates by diffusion welding, it is known to take these plates from sheets that have undergone a cold straightening process, called "skin pass" in English. Such sheets have a surface finish suitable for diffusion welding. However, during the diffusion welding of the plates, abnormal grain growth is frequently observed, resulting in the formation of grains with a size that can be more than ten times the average equivalent grain size of the alloy. Even if they are few in number, such large grains can drastically reduce the mechanical properties of the joint welded, such as creep resistance and fatigue strength. This abnormal grain growth can be attributed, in particular, to the existence, prior to welding, of local gradients of stored elastic energy between the grains. During cold straightening rolling, some grains more favorably oriented relative to the direction of stress deform, thus accumulating numerous crystalline defects and storing strain energy, while others, less favorably oriented, do not deform or deform very little. At the diffusion welding temperature, the grain boundaries separating deformed grains from undeformed grains are then likely to move much faster than the grain boundaries separating grains with similar deformation, resulting in strong heterogeneity in grain growth.

[0009] Furthermore, more generally, abnormal grain growth in an alloy can also be attributed to the reduction in the movement velocity of a grain boundary by fine second-phase particles that tend to block the movement of some of the boundaries, anchoring them. This phenomenon, known as the Zener-Smith effect, occurs when the particles are numerous and generally smaller than 1 µm. Anchored boundaries move little or not at all, while unanchored boundaries can move. Abnormal grain growth in an alloy can also have other causes, related, for example, to the presence of a texture or the dragging of solutes by grain boundaries.

[0010] For alloys exhibiting phase transformations, such as tool steels or martensitic steels, it is possible to regenerate a fine-grained microstructure by appropriate heat treatments. However, for other alloys, such as austenitic steels, nickel-based, aluminum-based, or copper-based alloys, regeneration of the microstructure by simple heat treatment is impossible.

[0011] Finally, at the interface between the contacting parts, it is desirable that the grain boundaries of the alloy in one part migrate into the other part, and vice versa, crossing the interface under the effect of the heating temperature. Such crossing contributes to the final homogenization of the microstructure of the welded assembly. However, poorly crossed interfaces may be observed, which exhibit reduced mechanical properties, for example, reduced creep resistance and / or fatigue resistance. The absence of crossing, or incomplete crossing, can have several causes. Since crossing depends on grain growth, the temperature and duration of the diffusion welding must therefore be sufficiently high.Furthermore, even if the conditions for grain growth are met, it can be observed that grain growth has little or no effect on the interface; that is, the penetration is incomplete or even absent. Such a phenomenon can be observed when the interface contains numerous obstacles, such as... small pores and inclusions originating from solid or gaseous contaminants present on the initial surfaces and / or in the welding atmosphere, or from particles forming spontaneously during welding from elements present in the material, are responsible for a type of Zener-Smith effect localized at the interface.

[0012] There is therefore a need for a production process for an assembly by diffusion welding that overcomes the aforementioned drawbacks. Summary of the invention

[0013] The invention relates to a process comprising the diffusion welding of a stack of parts, at least two of which are in contact and made of the same alloy, the alloy having:

[0014] - an absence of abnormal grain growth following a first reference heat treatment carried out under thermal conditions identical to the thermal conditions of diffusion welding, - a fully recrystallized microstructure with an average equivalent grain size of less than 50 µm following a second reference heat treatment which successively involves heating the alloy identical to the heating of the rise phase of the diffusion-welding process, and quenching, and - an average equivalent grain size at the end of the first reference heat treatment greater than or equal to 1.5 times the average equivalent grain size at the end of the second reference heat treatment.

[0015] Advantageously, the implementation of the process according to the invention results in an assembly with a homogeneous microstructure, in particular with a substantially unimodal grain size distribution, and in which the interfaces have been substantially overcome by the grain boundaries during welding. The assembly thus exhibits good mechanical properties, and in particular good resistance to chipping and good fatigue resistance. These mechanical properties result in particular from a non-excessive grain growth during diffusion welding.

[0016] The alloy exhibits an irreversible grain growth from the diffusion-welding process, that is to say, it has a matrix which does not undergo a phase transformation, during the descent phase or during a heat treatment subsequent to the diffusion-welding process, such that a microstructure with finer grains than at the end of the plateau phase can be obtained.

[0017] Preferably, the alloy is chosen from austenitic or ferritic stainless steels, nickel-based austenitic alloys, austenitic alloys based on iron and nickel, and copper alloys, for example a structurally hardening alloy.

[0018] The alloy may be a chromium-nickel austenitic stainless steel or a chromium-nickel-molybdenum austenitic stainless steel. Preferably, the alloy is an austenitic stainless steel not stabilized with titanium or niobium, preferably containing by mass less than 0.01% titanium and less than 0.02% niobium. In particular, the alloy may be selected from the stainless steel grades designated in standard NF EN 10088-1:2014 under numbers 1.43xy and 1.44xy, where x and y denote numeric characters.

[0019] The average equivalent grain size of the alloy can be determined, according to the composition of the alloy, in accordance with the standards in force such as the standards NF EN ISO 643:2020, NF EN ISO 2624, NF A04-503 and the ASTM-E112-13(2021).

[0020] The alloy according to the invention ensures that in the assembly obtained by diffusion welding, the average equivalent grain size does not exceed a value that is too high with regard to the expected mechanical properties of the assembly while allowing partial or complete crossing of the interfaces by the grain boundaries.

[0021] The two parts may each be in the form of a plate, and the interface of the assembly is defined by the large contacting faces of the plates. In particular, the thickness of at least one of the plates may be greater than or equal to 15 times the average equivalent grain size of the alloy after the second reference heat treatment.

[0022] In particular, the thickness of each plate can be between 0.2 mm and 5 mm. Advantageously, the process according to the invention makes it possible to manufacture an assembly having a sufficient number of grains throughout its thickness and thus exhibiting good mechanical properties. Each plate can have a length-to-thickness ratio greater than 100, or even greater than 1000.

[0023] In a preferred embodiment, the plate assembly obtained by the process according to the invention is intended to form in whole or in part a plate heat exchanger or a reactor-exchanger.

[0024] Furthermore, the alloy is characterized by an absence of abnormal grain growth, which can be easily observed, particularly after the first reference heat treatment. Specifically, the process may include a preliminary step of performing the first reference heat treatment on a sample of the alloy that is different from the parts to be assembled.

[0025] The "thermal conditions" implemented to perform the first reference heat treatment include at least the diffusion welding temperature and the duration of the dwell at the diffusion welding temperature. They may also include the rate of temperature rise to the diffusion welding temperature and / or rate of cooling from the diffusion welding temperature to a temperature significantly lower, for example 100°C, than that.

[0026] The first reference heat treatment is preferably carried out under atmospheric pressure. Preferably, during the first reference heat treatment, no external force is applied to the alloy. By "external force" is meant any force other than gravitational force. The first reference heat treatment and diffusion welding can be carried out in the same furnace.

[0027] The absence of abnormal grain growth in the alloy can be observed by examining an optical microscopy image of the alloy's microstructure containing at least 1000 grains. Abnormal growth is observed when the largest 1% of grains, by number, occupy more than 20% of the observed surface area. The absence of abnormal growth following the initial reference heat treatment allows us to anticipate that, after diffusion welding, the assembly will exhibit a homogeneous microstructure.

[0028] Furthermore, the alloy is characterized by an average equivalent grain size of less than 50 µm after the second reference heat treatment. In particular, the process may include a preliminary step of performing the second reference heat treatment on a sample made of the alloy and different from the parts to be assembled. The "thermal conditions" used to perform the second reference heat treatment include at least: a rapid temperature rise to the diffusion-welding temperature, no holding time, and quenching, for example, at a cooling rate of at least 10 °C / min.

[0029] The second reference heat treatment is preferably carried out under atmospheric pressure. Preferably, during the second reference heat treatment, no external force is applied to the alloy. The second reference heat treatment and diffusion welding can be carried out in the same furnace. The inventors have found that an average equivalent grain size of less than 50 µm following the second reference heat treatment allows for optimal penetration of the interface between the parts to be joined, while preventing the alloy from having an excessively large average equivalent grain size after diffusion welding.

[0030] After the second reference heat treatment, the alloy preferably has an average equivalent grain size of less than 20 pm.

[0031] As mentioned above, the absence of abnormal grain growth following the first reference heat treatment can be observed by carrying out the first reference heat treatment on a sample.

[0032] Alternatively, it can be observed from the results of a plurality of tests, each test being carried out without application of an external force on a sample made of the alloy, and involving maintaining the sample at a predetermined temperature and for a predetermined duration, the result of each of the tests including the observation of the absence of abnormal magnification.

[0033] Thus, by observing the absence of abnormal grain growth on all the tests, it can be determined that the alloy exhibits an absence of abnormal grain growth following the first reference heat treatment, without carrying out said first reference heat treatment.

[0034] For example, the temperature of at least one test may be lower than the diffusion welding temperature and the temperature of at least one test may be higher than the diffusion welding temperature and / or the duration of at least one test may be less than the duration of the diffusion welding plateau and the duration of at least one test may be higher than the duration of the diffusion welding plateau.

[0035] For example: - the test with the lowest temperature may have a temperature no more than 50 °C lower than the diffusion welding temperature and / or the test with the highest temperature may have a temperature no more than 50 °C higher than the diffusion welding temperature, and / or - the test with the shortest temperature holding time may have a duration of no more than 0.5 h less than the duration of the diffusion welding plateau and / or the test with the longest duration may have a duration of no more than 1 h greater than the duration of the diffusion welding plateau.

[0036] Furthermore, the average equivalent grain size following the first reference heat treatment and the average equivalent grain size following the second reference heat treatment can be measured by carrying out the first and second reference heat treatments respectively.

[0037] Alternatively, they can be determined from a model of the evolution of the average equivalent grain size, the model parameters being identified using experimental data measured during the tests. This makes it possible to eliminate the need for the second reference heat treatment.

[0038] For example, the model of evolution of the average equivalent grain size can be governed at least by the equation

[0039] [Math.l] Dj-D^Cf

[0040] in which: - Di is the initial average equivalent grain size, - / 7 / is the final average equivalent grain size after a test of duration ' at temperature T, and - C and n are parameters that depend on the temperature T.

[0041] Furthermore, the average equivalent grain size after the first reference heat treatment is greater than or equal to 1.5 times, or even greater than 2.5 times, the average equivalent grain size after the second reference heat treatment. The inventors have thus observed that normal grain growth can occur during the holding phase of the diffusion weld, which therefore ensures the crossing of the interface between the parts to be joined.

[0042] Preferably, after diffusion welding the assembly has an average equivalent grain size between 10 pm and 150 pm, preferably between 20 pm and 80 pm.

[0043] The holding time at the diffusion-welding temperature can be between 0.5 h and 6 h. The diffusion-welding temperature can, in particular, be greater than 0.7*Tf, where Tf is the absolute melting temperature of the alloy. The "absolute temperature" is the temperature referenced to absolute zero (0 K).

[0044] The diffusion welding technique is preferably chosen from hot isostatic compression and hot uniaxial compression. It involves compressing the assembly formed by the parts. The pressure applied to the parts during welding can be adjusted according to the alloy composition. In particular, it is sufficiently high to ensure the closure of the interface between the parts and to prevent the presence of pores in the assembly. Those skilled in the art know how to routinely select such a pressure.

[0045] Furthermore, the process preferably includes, prior to assembling the parts, cleaning at least one surface of each part and bringing the cleaned surfaces into contact. The cleaning can be carried out with a detergent, for example, chosen from alkaline detergents containing sodium hydroxide and acidic detergents, and / or a solvent, for example, chosen from oxygenated solvents, in particular acetone and / or alcohols, and halogenated solvents. The cleaned parts can then be rinsed and / or dried.

[0046] Preferably, the process further comprises degassing at least one cleaned surface, which may be rinsed and / or dried. Preferably, degassing is carried out by vacuuming the parts in contact. Degassing can be performed in situ, within the welding furnace when diffusion welding is performed by hot uniaxial compression, or by connecting a sealed container containing the parts to a vacuum pump when diffusion welding is performed by hot isostatic compression.

[0047] The assembly obtained by the process according to the invention may have a structure of equiaxed grains and in particular recrystallized.

[0048] Furthermore, the process may involve bringing one of the parts into contact with a third part made of another alloy which is linked to the assembly.

[0049] Finally, the invention relates to the use of an assembly obtained by the process according to the invention in a plate heat exchanger or in a reactor-exchanger.

[0050] The invention is illustrated hereafter by means of the following non-limiting examples and figures in which:

[0051] [Fig. 1] is a photograph acquired by optical microscopy of a polished surface of a sample of the steel of Example 1 after rolling and recrystallization annealing,

[0052] [Fig.2] is a photograph acquired by optical microscopy of a polished surface of a cross-section of an assembly obtained by an example of implementation of the process according to the invention,

[0053] [Fig.3] is a photograph acquired by optical microscopy of a polished surface of a cross-section of an assembly obtained by the non-invention process of Example 2. Example 1

[0054] An austenitic stainless steel X2CrNiMol7-12-2 is available, the composition of which in mass percentages is given in Table 1.

[0055] [Tables] Fe CN Mn Si SP Cr Ni 100% complement 0.021 0.072 1.71 0.19 <0.001 0.021 17.4 12.2 Mo Al Ti Nb V Ta BO 2.4 0.013 <0.001 0.003 0.046 <0.003 0.0002 0.0025

[0056] The steel was hot-rolled to a thickness of 10 mm and then cold-rolled to a thickness of 60% to form a 4 mm thick plate. It was then held at 950 °C for 10 minutes to ensure complete recrystallization. Following this treatment, the average equivalent grain size was 6.2 µm and no abnormal grain growth was observed ([Fig. 1]).

[0057] The steel was then studied for its grain growth behavior within the experimental temperature range of 975°C to 1100°C and for holding periods of up to 16 hours. No abnormal grain growth was observed during the tests under any of the conditions tested. It was therefore concluded that the steel is not subject to the phenomenon of abnormal grain growth within the experimental range considered, and that a homogeneous microstructure can be obtained after a heat treatment involving a heating to 1090°C followed by a 2-hour plateau, these conditions being part of the experimental range.

[0058] This study also made it possible to model the evolution of the average equivalent grain size of the steel, by approximating this size according to the equation £>2 p2 „ jans where D, and are the initial and final average equivalent grain sizes measured respectively before and after an exposure of duration f to temperature T, and C and n are parameters that depend on T, determined from the measurements of the study. This isothermal model can be used to estimate the grain growth during heating, by approximating the heating curve by a succession of small isothermal plateaus.

[0059] Using this model, it was determined that following heating to 1090°C, corresponding to the heating phase of a diffusion-welding cycle, without holding at 1090°C and followed by quenching (i.e., corresponding to the second reference treatment), the average equivalent grain size reaches 28.4 pm and is therefore less than 50 pm. It was further determined using the model that following a first reference heat treatment carried out under the thermal conditions of diffusion-welding and including a holding period of 2 hours at 1090°C, the average equivalent grain size is 46.2 pm, or 1.75 times larger than after the second reference heat treatment.

[0060] Subsequently, plates made of cold-rolled steel with a 60% reduction ratio, then recrystallized at 950°C (as described above), were cleaned, degassed, and then joined by hot isostatic compression diffusion welding at a welding temperature of 1090°C and under a pressure of 105 MPa for a duration of 2 hours. After cooling, as observed in [Fig. 2], following diffusion welding, the steel within the assembly exhibits a microstructure of equiaxed grains with an average equivalent size of 44.7 pm. The grain size distribution is substantially unimodal, and the assembly is free of abnormally sized grains. Furthermore, as observed in [Fig. 2], the interface between the two plates, the position of which before assembly is indicated by the arrows, is no longer visible and has been crossed by the grain boundaries of each of the plates during welding.Such an assembly exhibits excellent mechanical properties. Comparative Example 2

[0061] An austenitic stainless steel X2CrNiMol7-12-2 is available, the composition of which in mass percentages is given in Table 2.

[0062] [Tables2] Fe CN Mn Si s P Cr 100% complement 0.02 0.044 0.93 0.50 0.001 0.025 17.3 Ni Mo Al Ti Nb VBO 12.9 2.76 0.013 0.009 0.010 0.084 0.0013 0.0025

[0063] This steel was supplied in the form of 4 mm thick sheets after undergoing the following successive operations: hot rolling, cold rolling, recrystallization treatment at 1080°C followed by quenching, pickling, and application of a light cold rolling pass known as "skin pass". This last operation involves a total deformation of 1 to 3% and therefore results in the presence of local gradients of stored elastic energy. The average equivalent grain size is 30 µm.

[0064] Two plates made of the alloy thus slightly deformed are assembled under welding conditions identical to those of example 1.

[0065] After cooling, as observed in [Fig. 3], following diffusion welding, the steel exhibits a grain microstructure with a substantially bimodal size distribution. Abnormal grain growth is observed. This results in the presence of very large grains of several hundred microns, which impair the mechanical properties of the assembly.

Claims

Demands

1. A process comprising the diffusion welding of a stack of parts, of which at least two parts in contact are made of the same alloy, the alloy having: - an absence of abnormal grain growth following a first reference heat treatment carried out under thermal conditions identical to the thermal conditions of the diffusion welding, - a fully recrystallized microstructure with an average equivalent grain size of less than 50 pm following a second reference heat treatment which includes successively heating of the alloy identical to the heating of the rise phase of the diffusion welding, and quenching, and - an average equivalent grain size at the end of the first reference heat treatment greater than or equal to 1.5 times the average equivalent grain size at the end of the second reference heat treatment.

2. The method according to claim 1, the alloy being selected from austenitic or ferritic stainless steels, nickel-based austenitic alloys, iron-nickel-based austenitic alloys, and copper alloys, for example a structurally hardening alloy.

3. The process according to claim 2, the alloy being an austenitic stainless steel not stabilized with titanium or niobium, preferably comprising by mass less than 0.01% titanium and less than 0.02% niobium.

4. Process according to any one of the preceding claims, the alloy having an average equivalent grain size of less than 20 pm after the second reference heat treatment.

5. Process according to any one of the preceding claims, the alloy having an average equivalent grain size at the end of the first reference heat treatment greater than or equal to 2.5 times the average equivalent grain size at the end of the second reference heat treatment.

6. A process according to any one of the preceding claims, the absence of abnormal grain growth following the first reference heat treatment being observed by carrying out the first reference heat treatment.

7. A method according to any one of claims 1 to 6, the absence of abnormal grain growth following the first reference heat treatment being observed from the results of a plurality of tests, each test being carried out without application of an external force on a sample made of the alloy, and including maintaining the sample at a predetermined temperature and for a predetermined duration, the result of each of the tests including the finding of the absence of abnormal growth.

8. A method according to any one of the preceding claims, wherein the average equivalent grain size following the first reference heat treatment and the average equivalent grain size following the second reference heat treatment are measured by carrying out the first and second reference heat treatments respectively.

9. Method according to claim 7, the average equivalent grain size following the first reference heat treatment and the average equivalent grain size following the second heat treatment being determined from an evolution model of the average equivalent grain size, the parameters of the model being identified by means of experimental data measured during the tests.

10. Method according to claim 9, the model of evolution of the average equivalent grain size being governed at least by the equation: Dj -Dj=C t" in which: - Dj is the initial average equivalent grain size, - ^ / is the final average equivalent grain size after carrying out a test of a duration f at temperature T, and - C and n are parameters which depend on the temperature T.

11. A method according to any one of the preceding claims, the welding-diffusion technique being selected from hot isostatic compression and hot uniaxial compression.

12. A method according to any one of the preceding claims, comprising prior to the assembly of the parts, cleaning at least one surface of each part and bringing the cleaned surfaces into contact, and preferably degassing at least one cleaned surface, and where appropriate rinsing and / or drying.

13. A method according to any one of the preceding claims, the two parts each having the form of a plate and the interface of the assembly being defined by the large contacting faces of the plates.

14. The method according to the preceding claim, the thickness of at least one of the plates being greater than or equal to 15 times the average equivalent grain size of the alloy after the second reference heat treatment.

15. Use of an assembly obtained by the process according to any one of the preceding claims in a plate heat exchanger or in a reactor-exchanger.