A method for vacuum sputtering and heat treatment of high-precision thin-film resistors

Abstract

This invention relates to the field of electronic component manufacturing technology, and discloses a method for vacuum sputtering and heat treatment of high-precision thin-film resistors. The method includes depositing a nickel-chromium alloy thin film in pure argon gas, followed by in-situ nitriding with nitrogen gas during the later stages of deposition. The substrate is then transferred to an annealing furnace, where it is first held at a constant temperature under pure vacuum to release internal stress in the thin film. A premixed hydrogen-argon forming gas is then introduced to further increase the temperature, reducing and cleaning grain boundary oxides. After the heating is complete, high-purity argon gas is used for gas phase buffering and purging to reduce the total gas pressure. A premixed oxygen-argon mixture is then introduced, and a chromium trioxide layer is formed at a constant temperature under a set low oxygen partial pressure. Finally, high-pressure argon gas is introduced for pressurization and cooling. This invention, through the combination of multi-stage sputtering and atmosphere-controlled annealing, reduces the temperature coefficient of resistance of the thin film, improves the structural stability and long-term service reliability of the resistor, and solves the problem of thin-film resistance drift.

Description

Technical Field

[0001] This invention relates to the field of electronic component manufacturing technology, specifically to a method for vacuum sputtering and heat treatment of high-precision thin-film resistors. Background Technology

[0002] High-precision thin-film resistors are widely used in electronic measurement and precision instruments. Their core performance indicators mainly include the temperature coefficient of resistance and the resistance stability under long-term service. Currently, the manufacturing of such resistors generally adopts vacuum magnetron sputtering technology to deposit nickel-chromium alloy thin films on the substrate, and then uses heat treatment processes to solidify the electrical properties of the thin film.

[0003] Existing thin film sputtering and heat treatment methods still have some limitations in actual production. During the sputtering process, residual internal stress accumulates inside the film and lattice defects are generated. If the subsequent heat treatment process cannot eliminate these stresses and defects, the resistor is prone to microstructural changes due to the slow release of internal stress during use, which in turn causes resistance drift. At the same time, the existing high-temperature annealing process usually has relatively simple control over the chamber atmosphere. The trace amount of residual oxygen in the furnace can easily cause disordered oxidation at the grain boundaries of the film. These primary oxides are generally not stable enough and will interfere with the heat treatment recrystallization process of the alloy grains, resulting in a larger temperature coefficient of thin film resistance.

[0004] Furthermore, when switching between reducing and oxidizing atmospheres during the heat treatment stage, conventional processes often directly introduce gases of different properties. Due to the lack of a phase change buffer process, this can lead to uneven reaction rates in local areas of the substrate. This directly affects the density of the surface passivation layer and the consistency of the overall film structure, making it difficult for the final product to meet the high-precision stability requirements. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for vacuum sputtering and heat treatment of high-precision thin-film resistors, which solves the technical problems of excessively large temperature coefficient of resistance and resistance drift during long-term service caused by internal residual stress and disordered oxidation at grain boundaries in the manufacturing process of existing high-precision thin-film resistors.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for vacuum sputtering and heat treatment of a high-precision thin-film resistor, comprising the following steps: S1. Load the single-crystal silicon substrate into the magnetron sputtering chamber for backsputter cleaning; then use the nickel-chromium alloy target as the sputtering source to deposit a thin film in a pure high-purity argon atmosphere. When the deposition thickness reaches the first set thickness of the target total thickness, high-purity nitrogen is dynamically introduced, and sputtering continues in a mixed plasma atmosphere to complete the remaining film thickness deposition, thus obtaining a single-crystal silicon substrate with a thin film attached. S2. High-purity hydrogen and high-purity argon are simultaneously introduced into a sealed mixing tank for physical mixing to obtain a shaped gas. S3. The single-crystal silicon substrate with the thin film is moved into the vacuum annealing furnace and evacuated to the set background vacuum level. The temperature is raised to the first set temperature and held at a constant temperature under pure vacuum. After the holding temperature is completed, forming gas is introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure in the furnace under the reducing pressure holding state of the set pressure and continue to raise the temperature. S4. High-purity oxygen and high-purity argon are simultaneously introduced into a closed mixing tank for physical mixing to obtain an oxidizing mixture. S5. When the furnace temperature reaches the second set temperature, cut off the forming gas inlet and introduce high-purity argon for purging. At the same time, adjust the exhaust valve to reduce the total gas pressure in the furnace. After purging, cut off the high-purity argon inlet and introduce an oxidizing mixed gas to establish the oxygen partial pressure in the furnace at the set partial pressure value and perform constant temperature holding treatment. S6. After the holding treatment is completed, cut off the oxidizing mixed gas inlet and introduce high-purity argon into the vacuum annealing furnace for pressurization and cooling. Open the furnace and take out the high-precision thin film resistor.

[0007] By adopting the above technical solution, this invention combines multi-stage sputtering with atmosphere-controlled annealing, supplemented by gas distribution and partial pressure control, thereby achieving the goal of reducing the temperature coefficient of thin film resistance and improving the stability of the thin film structure. Throughout the entire process, the microscopic mechanism of the above technical solution is manifested as follows: For the thin film deposition process, S1 deposits a pure nickel-chromium alloy in a pure argon atmosphere to form a conductive host layer. Nitrogen gas is then introduced during the subsequent deposition process. The mixed plasma environment within the chamber induces a nitriding reaction on the film surface, generating a nickel-chromium nitride layer in situ. This layer acts as a physical barrier, inhibiting the diffusion of oxygen atoms into the deeper layers of the film during subsequent high-temperature processing.

[0008] After the thin film is prepared, the key is to deal with the internal defects and stress. S3 uses constant temperature in a pure vacuum environment to release the internal stress accumulated during the film deposition and repair the lattice defects. Then, forming gas is introduced and the temperature is continued to rise. During the heating stage, hydrogen exhibits reduction properties and reacts with the initial metal oxides on the film surface and at the grain boundaries caused by trace amounts of oxygen, thereby purifying the grain boundaries.

[0009] When the above reduction and crystallization processes are completed, the process enters the gas phase replacement and passivation stage. In S5, the forming gas is cut off and high-purity argon is introduced for gas phase buffer scavenging. This operation can remove residual hydrogen in the chamber, avoid the risk of direct oxygen introduction and reaction, and at the same time reduce the total gas pressure to provide airflow space for subsequent passivation. After the oxidizing mixed gas is introduced, isothermal passivation is carried out at the set partial pressure value. Trace amounts of oxygen undergo selective oxidation reaction with chromium in the nickel-chromium alloy, generating a chromium trioxide passivation layer on the surface of the thin film and at the shallow grain boundaries.

[0010] In the final stage, S6 is filled with high-pressure, high-purity argon gas for pressurization and cooling, in order to terminate the surface oxidation process and accelerate the cooling of the substrate, preventing abnormal grain growth.

[0011] Preferably, the specific operation of the cleaning process in S1 is as follows: a single-crystal silicon substrate with a thermally oxidized silicon insulating layer on its surface is loaded into a magnetron sputtering chamber, the base vacuum of the magnetron sputtering chamber is evacuated to a set sputtering base vacuum, high-purity argon gas is introduced to maintain the absolute pressure in the chamber at 0.5 Pa to 1.5 Pa, and the radio frequency power supply is started to perform reverse sputtering cleaning on the single-crystal silicon substrate for 5 to 15 minutes.

[0012] By adopting the above technical solution, radio frequency plasma bombardment of the substrate surface can remove residual adsorbed impurities, increase the surface roughness of the substrate, and thus enhance the physical bonding force between the single crystal silicon substrate and the nickel-chromium alloy film.

[0013] Preferably, the in-situ nitriding deposition process in S1 is specifically operated as follows: after the cleaning process, a nickel-chromium alloy target composed of 40 to 80 parts by weight of nickel and 20 to 60 parts by weight of chromium is used as the sputtering source, with a sputtering power density of 2.0 W / cm². 2 Up to 8.0 W / cm 2 A thin film is deposited under a pure high-purity argon atmosphere of 0.5 Pa to 1.5 Pa. When the deposition thickness reaches 85% to 95% of the target total thickness, high-purity nitrogen is dynamically introduced so that its volume flow rate accounts for 3% to 8% of the total inlet flow rate. The remaining 5% to 15% of the film thickness is deposited by sputtering in a mixed plasma atmosphere, resulting in a single-crystal silicon substrate with a thin film.

[0014] By adopting the above technical solution, the nitrogen content is controlled by limiting the target material composition and airflow distribution ratio, so that the surface nitride layer has a suitable density. This configuration avoids the problem of nonlinear increase in film resistivity caused by excessive nitriding layer.

[0015] Preferably, the specific operation of gas mixing in S2 is as follows: under the conditions of setting the mixing temperature and setting the mixing pressure, the gas mass flow meter is turned on, the inlet flow rate of high-purity hydrogen is set to 30 ml / min to 100 ml / min, and the inlet flow rate of high-purity argon is set to 900 ml / min to 970 ml / min. The above two gases are simultaneously introduced into a sealed mixing tank with a set volume, and the gas is continuously introduced for a set mixing time to obtain a forming gas with a hydrogen volume ratio of 3% to 10%. This process is a physical gas mixing and no chemical reaction occurs.

[0016] By adopting the above technical solution, the gas is physically mixed in a closed mixing tank in advance, which can obtain a reducing gas with uniform composition. At the same time, the volume ratio of hydrogen is controlled in a low concentration range, which can ensure the safety of the production process while ensuring the reducing ability.

[0017] Preferably, the specific operation of the first-order oxygen-free annealing in S3 is as follows: the single crystal silicon substrate with the thin film is moved into the vacuum annealing furnace, the background vacuum degree in the furnace is evacuated to the set annealing background vacuum degree, the temperature is raised to 220°C to 280°C at a heating rate of 3°C / min to 8°C / min, and held at a constant temperature under pure vacuum for 1.5 hours to 3.0 hours.

[0018] By adopting the above technical solution, setting a slower heating rate can reduce the probability of substrate deformation due to uneven thermal expansion. On this basis, by using pure vacuum annealing conditions in the medium temperature range, point defects inside the grains can be promoted to migrate to the grain boundaries and be annihilated.

[0019] Preferably, the subsequent forming gas pressure holding and heating operation in S3 is as follows: after the heat holding is completed, the furnace temperature is kept constant, and the forming gas prepared in S2 is introduced into the vacuum annealing furnace to establish and stabilize the absolute gas pressure in the furnace at 400Pa to 900Pa. Under this reducing pressure holding state, the furnace temperature is raised from 220°C to 280°C to 350°C to 410°C at a heating rate of 2°C / min to 5°C / min.

[0020] By adopting the above technical solution, and by applying a reducing atmosphere during the stage of raising the temperature, the activity of residual oxygen on the film surface is reduced. This treatment mechanism avoids uncontrolled oxidation behavior of the film at high temperatures.

[0021] Preferably, the specific operation of gas mixing in S4 is as follows: under the conditions of setting the mixing temperature and setting the mixing pressure, the gas mass flow meter is turned on, the inlet flow rate of high-purity oxygen is set to 50 ml / min to 200 ml / min, and the inlet flow rate of high-purity argon is set to 800 ml / min to 950 ml / min. The above two gases are simultaneously introduced into a sealed mixing tank with a set volume, and the gas is continuously introduced for a set mixing time to obtain an oxidizing mixture with an oxygen volume ratio of 5% to 20%. This process is a physical gas mixing and no chemical reaction occurs.

[0022] By adopting the above technical solution, oxygen and argon are pre-mixed to avoid localized uneven oxidation caused by directly introducing high-concentration oxygen into the high-temperature furnace chamber, thus ensuring the uniformity of the subsequent passivation layer growth thickness.

[0023] Preferably, the specific operation of the gas phase buffer purging in S5 is as follows: when the furnace temperature reaches 350°C to 410°C, the forming gas inlet is cut off, and high-purity argon gas is introduced instantly for purging for 1 to 3 minutes. At the same time, the exhaust valve is adjusted to reduce the total gas pressure in the furnace to 20 Pa to 80 Pa.

[0024] By adopting the above technical solution, high-purity argon gas is used to rapidly replace the reducing atmosphere in the furnace, while reducing the absolute pressure inside the furnace. This purging operation provides process space for the establishment of a low-oxygen partial pressure environment.

[0025] Preferably, the subsequent isothermal oxidation passivation operation in S5 is as follows: after the cleaning is completed, the high-purity argon gas inlet is cut off within 10 to 30 seconds, and the oxidizing mixed gas prepared in S4 is introduced to accurately establish the oxygen partial pressure in the furnace at 8 Pa to 20 Pa. Then, the furnace temperature is maintained at 350°C to 410°C and the oxygen partial pressure is maintained at 8 Pa to 20 Pa for isothermal heat treatment for 1.0 hour to 2.5 hours.

[0026] By adopting the above technical solution, under the set low oxygen partial pressure, the diffusion rate of chromium atoms on the surface of the thin film and the oxidation reaction rate can form a dynamic balance, thereby generating a dense passivation layer.

[0027] Preferably, the specific operation of S6 is as follows: after the heat preservation is completed, the oxidizing mixed gas inlet is cut off, high-purity argon gas is introduced into the vacuum annealing furnace until the internal pressure reaches 80kPa to 110kPa, the heating system is turned off and the single crystal silicon substrate with thin film is cooled to the set furnace exit temperature at a cooling rate of 15℃ / min to 30℃ / min, the furnace is opened and the substrate is taken out to obtain a high-precision thin film resistor.

[0028] By adopting the above technical solution, high-pressure argon gas is introduced to accelerate the removal of heat from the film surface and increase the cooling rate, thereby solidifying the microstructure of the film.

[0029] This invention provides a method for vacuum sputtering and heat treatment of high-precision thin-film resistors. It offers the following advantages: 1. The present invention forms a physical barrier layer on the surface of the thin film through the in-situ nitriding deposition process of S1, and combines it with the low oxygen partial pressure isothermal oxidation treatment of S5 to generate a dense chromium trioxide passivation layer on the surface of the thin film. The barrier layer and the passivation layer work together to suppress the resistance drift caused by environmental factors during the use of the resistor, thereby improving the long-term stability and reliability of the resistor.

[0030] 2. The present invention employs a two-stage annealing process in S3, which first involves pure vacuum heat preservation followed by heating with forming gas. The former releases the internal stress accumulated during the deposition process, while the latter utilizes the reducing properties of hydrogen to remove unstable metal oxides at the grain boundaries and promotes the recrystallization of the film. The resulting film structure is more regular and the defect density is reduced, thereby significantly reducing the temperature coefficient of resistance of the resistor.

[0031] 3. This invention controls the process atmosphere in stages, such as preparing a uniformly mixed gas in S2 and S4 in advance, and setting a gas phase buffer sweeping step in S5. These processing mechanisms avoid local uncontrolled reactions caused by incomplete atmosphere switching or uneven gas concentration, ensuring the controllability and consistency of the nitriding and oxidation passivation processes, thereby improving the yield and performance uniformity of the product. Attached Figure Description

[0032] Figure 1 The graphs show the changes in macroscopic residual stress and mass of the thin film, where (a) is the evolution trend of macroscopic residual stress at each process stage; and (b) is the change in micro-mass after first-stage annealing. Figure 2 The images show the in-situ thermal and mass spectrometry dynamic monitoring data, where (a) is a comparison of the peak concentration of nitrogen gas escaping during the annealing process; and (b) is a distribution of the transient temperature rise on the substrate surface. Figure 3 Thermogravimetric analysis curve of the thin film of the present invention in an extreme high temperature oxygen-enriched environment of 500℃. Figure 4 The following are the electrical stability diagrams of the thin film under high temperature aging, where (a) is the time evolution curve of the sheet resistance drift rate and a magnified view of the part; (b) is a comparison diagram of the temperature coefficient of resistance before and after aging. Figure 5 A comparison chart of the self-corrosion potentials of each sample measured in 3.5% NaCl solution; Figure 6 Here is a comparison chart of the corresponding self-corrosion current densities; Figure 7 This is a graph showing the relationship between in-plane residual stress and critical peel load. Figure 8 This is the real-time response curve of the friction coefficient of a typical sample in a continuous variable load scratch test. Detailed Implementation

[0033] The present invention will be further described below with reference to specific embodiments. The specific embodiments of the present invention are for illustrative purposes only and are not intended to limit the invention. Those skilled in the art should understand that conventional substitutions or adjustments to the process steps, reaction conditions, or parameters of the present invention without departing from the core ideas of the invention are all within the scope of protection of the present invention.

[0034] The main raw materials and reagents used in the following examples and comparative examples have the following sources and specifications. Reagents not specifically mentioned are all commercially available analytical grade or higher grade products.

[0035] The nickel-chromium alloy sputtering target is a commercially available industrial-grade high-purity sputtering target, in which the mass fraction of nickel is 40% to 80% (CAS No. 7440-02-0). The chromium content is 20% to 60% by mass (CAS No. 7440-47-3), and the sum of the nickel and chromium mass fractions is 100%, with an overall purity greater than or equal to 99.99%.

[0036] The substrate is a commercially available alumina ceramic substrate with a purity greater than or equal to 99.6% (CAS No. 1344-28-1), or an electronic-grade single-crystal silicon substrate with a thermally heated silicon oxide insulating layer with a thickness of 200 nanometers to 500 nanometers.

[0037] The experiment used high-purity argon gas (CAS number 7440-37-1). High-purity nitrogen (CAS No. 7727-37-9); High-purity hydrogen (CAS No. 1333-74-0); High-purity oxygen (CAS No. 7782-44-7); its gaseous volume purity is greater than or equal to 99.999%.

[0038] Example 1: This example provides a method for vacuum sputtering and heat treatment of a high-precision thin-film resistor, including the following steps: S1. Load a single-crystal silicon substrate with a thermally oxidized silicon insulating layer onto the magnetron sputtering chamber, and evacuate the base vacuum of the magnetron sputtering chamber to 4.0 × 10⁻⁶. -4 The pressure inside the chamber was maintained at 1.5 Pa by introducing high-purity argon gas. The radio frequency power supply was then activated to perform reverse sputtering cleaning of the single-crystal silicon substrate for 15 minutes. Subsequently, a nickel-chromium alloy target composed of 80 parts by weight of nickel and 20 parts by weight of chromium was used as the sputtering source, with a sputtering power density of 8.0 W / cm². 2 A thin film was deposited under a pure high-purity argon atmosphere of 1.5 Pa. When the deposition thickness reached 95% of the target total thickness, high-purity nitrogen was dynamically introduced so that its volume flow rate accounted for 8% of the total inlet flow rate. The remaining 5% of the film thickness was deposited by sputtering in a mixed plasma atmosphere, resulting in a single-crystal silicon substrate with a thin film. S2. Under the conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity hydrogen to 100 ml / min, and set the inlet flow rate of high-purity argon to 900 ml / min. Simultaneously introduce the above two gas streams into a 5-liter sealed mixing tank and continue to circulate the gas for 20 minutes to obtain a forming gas with a hydrogen volume ratio of 10%. This process is a physical gas mixing and no chemical reaction occurs. S3. Transfer the single-crystal silicon substrate with the thin film to a vacuum annealing furnace, and evacuate the furnace to a background vacuum of 0.8 × 10⁻⁶. - 3 Pa, the temperature was raised to 280°C at a heating rate of 8°C / min and held at a constant temperature under pure vacuum for 3.0 hours; after the holding period, the furnace temperature was kept constant and the forming gas prepared by S2 was introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure in the furnace at 900Pa. Under this reducing pressure holding state, the furnace temperature was raised from 280°C to 410°C at a heating rate of 5°C / min. S4. Under conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity oxygen to 200 ml / min, and the inlet flow rate of high-purity argon to 800 ml / min. Simultaneously introduce both gases into a 5-liter sealed mixing tank and continue purging for 20 minutes to obtain an oxidizing mixture with an oxygen volume ratio of 20%. This process is a physical gas mixing and does not involve a chemical reaction.

[0039] S5. When the furnace temperature reaches 410℃, cut off the forming gas inlet and instantly introduce high-purity argon gas for purging for 3 minutes. At the same time, adjust the exhaust valve to smoothly transition the total gas pressure in the furnace to 80Pa. After purging, cut off the high-purity argon gas inlet within 30 seconds and introduce the oxidizing mixed gas prepared in S4 to accurately establish the oxygen partial pressure in the furnace at 20Pa. Then, maintain the furnace temperature at 410℃ and maintain the oxygen partial pressure at 20Pa for constant temperature heat preservation treatment for 2.5 hours. S6. After the heat preservation is completed, the oxidizing mixed gas inlet is cut off, and high-purity argon gas is introduced into the vacuum annealing furnace until the internal pressure reaches 110 kPa. The heating system is turned off and the single crystal silicon substrate with thin film is cooled to 90°C at a cooling rate of 30°C / min. The furnace is opened and the substrate is taken out to obtain a high-precision thin film resistor.

[0040] Example 2: This example provides a method for vacuum sputtering and heat treatment of a high-precision thin-film resistor, including the following steps: S1. Load a single-crystal silicon substrate with a thermally oxidized silicon insulating layer onto the magnetron sputtering chamber, and evacuate the base vacuum of the magnetron sputtering chamber to 4.0 × 10⁻⁶. -4 The pressure inside the chamber was maintained at 1.0 Pa by introducing high-purity argon gas. The radio frequency power supply was then activated to perform reverse sputtering cleaning of the single-crystal silicon substrate for 10 minutes. Subsequently, a nickel-chromium alloy target composed of 60 parts by weight of nickel and 40 parts by weight of chromium was used as the sputtering source, with a sputtering power density of 5.0 W / cm². 2A thin film was deposited under a pure high-purity argon atmosphere of 1.0 Pa. When the deposition thickness reached 90% of the target total thickness, high-purity nitrogen was dynamically introduced so that its volume flow rate accounted for 5.5% of the total inlet flow rate. The remaining 10% of the film thickness was deposited by sputtering in a mixed plasma atmosphere, resulting in a single-crystal silicon substrate with a thin film. S2. Under the conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity hydrogen to 60 ml / min, and set the inlet flow rate of high-purity argon to 940 ml / min. Simultaneously introduce the above two gases into a 5-liter sealed mixing tank and continue to circulate the gas for 20 minutes to obtain a forming gas with a hydrogen volume ratio of 6%. This process is a physical gas mixing and no chemical reaction occurs. S3. Transfer the single-crystal silicon substrate with the thin film to a vacuum annealing furnace, and evacuate the furnace to a background vacuum of 0.8 × 10⁻⁶. - 3 Pa, the temperature was raised to 250℃ at a heating rate of 5.5℃ / min and held at a constant temperature under pure vacuum for 2.25 hours; after the holding period, the furnace temperature was kept constant and the forming gas prepared by S2 was introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure in the furnace at 650Pa. Under this reducing pressure holding state, the furnace temperature was raised from 250℃ to 380℃ at a heating rate of 3.5℃ / min. S4. Under the conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity oxygen to 120 ml / min, and set the inlet flow rate of high-purity argon to 880 ml / min. Simultaneously introduce the above two gases into a closed mixing tank with a volume of 5 liters and continue to circulate the gas for 20 minutes to obtain an oxidizing mixture with an oxygen volume ratio of 12%. This process is a physical gas mixing and no chemical reaction occurs. S5. When the furnace temperature reaches 380℃, cut off the forming gas inlet and instantly introduce high-purity argon gas for 2 minutes to purge. At the same time, adjust the exhaust valve to smoothly transition the total gas pressure in the furnace to 50Pa. After purging, cut off the high-purity argon gas inlet within 20 seconds and introduce the oxidizing mixed gas prepared in S4 to accurately establish the oxygen partial pressure in the furnace at 14Pa. Then, maintain the furnace temperature at 380℃ and maintain the oxygen partial pressure at 14Pa for constant temperature heat preservation treatment for 1.75 hours. S6. After the heat preservation is completed, the oxidizing mixed gas inlet is cut off, and high-purity argon gas is introduced into the vacuum annealing furnace until the internal pressure reaches 95 kPa. The heating system is turned off and the single crystal silicon substrate with thin film is cooled to 90°C at a cooling rate of 22.5°C / min. The furnace is opened and the substrate is taken out to obtain a high-precision thin film resistor.

[0041] Example 3: This example provides a method for vacuum sputtering and heat treatment of a high-precision thin-film resistor, including the following steps: S1. Load a single-crystal silicon substrate with a thermally oxidized silicon insulating layer onto the magnetron sputtering chamber, and evacuate the base vacuum of the magnetron sputtering chamber to 4.0 × 10⁻⁶. -4 The pressure inside the chamber was maintained at 0.5 Pa by introducing high-purity argon gas. The radio frequency power supply was then activated to perform back-sputter cleaning of the single-crystal silicon substrate for 5 minutes. Subsequently, a nickel-chromium alloy target composed of 40 parts by weight of nickel and 60 parts by weight of chromium was used as the sputtering source, with a sputtering power density of 2.0 W / cm². 2 A thin film was deposited under a pure high-purity argon atmosphere of 0.5 Pa. When the deposition thickness reached 85% of the target total thickness, high-purity nitrogen was dynamically introduced so that its volume flow rate accounted for 3% of the total inlet flow rate. The remaining 15% of the film thickness was deposited by sputtering in a mixed plasma atmosphere, resulting in a single-crystal silicon substrate with a thin film. S2. Under conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity hydrogen to 30 ml / min, and the inlet flow rate of high-purity argon to 970 ml / min. Simultaneously introduce both gases into a 5-liter sealed mixing tank and continue purging for 20 minutes to obtain a shaped gas with a hydrogen volume ratio of 3%. This process is a physical gas mixing and does not involve any chemical reaction. S3. Transfer the single-crystal silicon substrate with the thin film to a vacuum annealing furnace, and evacuate the furnace to a background vacuum of 0.8 × 10⁻⁶. - 3 Pa, the temperature was raised to 220°C at a heating rate of 3°C / min and held at a constant temperature under pure vacuum for 1.5 hours; after the holding time, the furnace temperature was kept constant and the forming gas prepared by S2 was introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure in the furnace at 400Pa. Under this reducing pressure holding state, the furnace temperature was raised from 220°C to 350°C at a heating rate of 2°C / min. S4. Under the conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity oxygen to 50 ml / min, and set the inlet flow rate of high-purity argon to 950 ml / min. Simultaneously introduce the above two gases into a closed mixing tank with a volume of 5 liters and continue to circulate the gas for 20 minutes to obtain an oxidizing mixture with an oxygen volume ratio of 5%. This process is a physical gas mixing and no chemical reaction occurs. S5. When the furnace temperature reaches 350℃, cut off the forming gas inlet and instantly introduce high-purity argon gas for 1 minute to purge. At the same time, adjust the exhaust valve to smoothly transition the total gas pressure in the furnace to 20Pa. After purging, cut off the high-purity argon gas inlet within 10 seconds and introduce the oxidizing mixed gas prepared in S4 to accurately establish the oxygen partial pressure in the furnace at 8Pa. Then, maintain the furnace temperature at 350℃ and maintain the oxygen partial pressure at 8Pa for constant temperature heat preservation treatment for 1.0 hour. S6. After the heat preservation is completed, the oxidizing mixed gas inlet is cut off, and high-purity argon gas is introduced into the vacuum annealing furnace until the internal pressure reaches 80 kPa. The heating system is turned off and the single crystal silicon substrate with thin film is cooled to 90°C at a cooling rate of 15°C / min. The furnace is opened and the substrate is taken out to obtain a high-precision thin film resistor.

[0042] Example 4: This example provides a method for vacuum sputtering and heat treatment of a high-precision thin-film resistor, including the following steps: S1. Load a single-crystal silicon substrate with a thermally oxidized silicon insulating layer onto the magnetron sputtering chamber, and evacuate the base vacuum of the magnetron sputtering chamber to 4.0 × 10⁻⁶. -4 The pressure inside the chamber was maintained at 1.2 Pa by introducing high-purity argon gas. The RF power supply was then activated to perform reverse sputtering cleaning of the single-crystal silicon substrate for 12 minutes. Subsequently, a nickel-chromium alloy target composed of 70 parts by weight of nickel and 30 parts by weight of chromium was used as the sputtering source, with a sputtering power density of 6.0 W / cm². 2 A thin film was deposited under a pure high-purity argon atmosphere of 1.2 Pa. When the deposition thickness reached 92% of the target total thickness, high-purity nitrogen was dynamically introduced so that its volume flow rate accounted for 7% of the total inlet flow rate. The remaining 8% of the film thickness was deposited by sputtering in a mixed plasma atmosphere, resulting in a single-crystal silicon substrate with a thin film. S2. Under conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity hydrogen to 80 ml / min, and the inlet flow rate of high-purity argon to 920 ml / min. Simultaneously introduce both gases into a 5-liter sealed mixing tank and continue purging for 20 minutes to obtain a forming gas with a hydrogen volume ratio of 8%. This process is a physical gas mixing and does not involve any chemical reaction. S3. The single-crystal silicon substrate with the thin film is transferred into a vacuum annealing furnace. The base vacuum in the furnace is evacuated to 0.8×10-3 Pa. The temperature is raised to 260°C at a heating rate of 6°C / min and held at a constant temperature under pure vacuum for 2.5 hours. After the holding period, the furnace temperature is kept constant. The forming gas prepared in S2 is introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure in the furnace at 800 Pa. Under this reducing pressure holding condition, the furnace temperature is raised from 260°C to 400°C at a heating rate of 4°C / min. S4. Under conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity oxygen to 150 ml / min, and the inlet flow rate of high-purity argon to 850 ml / min. Simultaneously introduce both gases into a 5-liter sealed mixing tank and continue purging for 20 minutes to obtain an oxidizing mixture with an oxygen volume ratio of 15%. This process is a physical gas mixing and does not involve a chemical reaction. S5. When the furnace temperature reaches 400℃, cut off the forming gas inlet and instantly introduce high-purity argon gas for purging for 2.5 minutes. At the same time, adjust the exhaust valve to smoothly transition the total gas pressure in the furnace to 60Pa. After purging, cut off the high-purity argon gas inlet within 25 seconds and introduce the oxidizing mixed gas prepared in S4 to accurately establish the oxygen partial pressure in the furnace at 18Pa. Then, maintain the furnace temperature at 400℃ and maintain the oxygen partial pressure at 18Pa for constant temperature heat preservation treatment for 2.0 hours. S6. After the heat preservation is completed, the oxidizing mixed gas inlet is cut off, and high-purity argon gas is introduced into the vacuum annealing furnace until the internal pressure reaches 105 kPa. The heating system is turned off and the single crystal silicon substrate with thin film is cooled to 90°C at a cooling rate of 25°C / min. The furnace is opened and the substrate is taken out to obtain a high-precision thin film resistor.

[0043] Example 5: This example provides a method for vacuum sputtering and heat treatment of a high-precision thin-film resistor, including the following steps: S1. Load a single-crystal silicon substrate with a thermally oxidized silicon insulating layer onto the magnetron sputtering chamber, and evacuate the base vacuum of the magnetron sputtering chamber to 4.0 × 10⁻⁶. -4 The pressure inside the chamber was maintained at 0.8 Pa by introducing high-purity argon gas. The radio frequency power supply was then activated to perform back-sputter cleaning of the single-crystal silicon substrate for 8 minutes. Subsequently, a nickel-chromium alloy target composed of 50 parts by weight of nickel and 50 parts by weight of chromium was used as the sputtering source, with a sputtering power density of 4.0 W / cm². 2 A thin film was deposited under a pure high-purity argon atmosphere of 0.8 Pa. When the deposition thickness reached 88% of the target total thickness, high-purity nitrogen was dynamically introduced so that its volume flow rate accounted for 4% of the total inlet flow rate. The remaining 12% of the film thickness was deposited by sputtering in a mixed plasma atmosphere, resulting in a single-crystal silicon substrate with a thin film. S2. Under conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter and set the inlet flow rate of high-purity hydrogen to 50 mL / min and the inlet flow rate of high-purity argon to 950 mL / min. Simultaneously introduce both gases into a 5-liter sealed mixing tank and continue purging for 20 minutes to obtain a shaped gas with a hydrogen volume percentage of 5%. This process is a physical gas mixing process and does not involve any chemical reaction. S3. Transfer the single-crystal silicon substrate with the thin film to a vacuum annealing furnace, and evacuate the furnace to a background vacuum of 0.8 × 10⁻⁶. - 3 Pa, the temperature was raised to 240°C at a heating rate of 4°C / min and held at a constant temperature under pure vacuum for 2.0 hours; after the holding time, the furnace temperature was kept constant and the forming gas prepared by S2 was introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure in the furnace at 500Pa. Under this reducing pressure holding state, the furnace temperature was raised from 240°C to 360°C at a heating rate of 3°C / min. S4. Under conditions of 25℃ and 0.2 MPa pressure, turn on the gas mass flow meter, set the inlet flow rate of high-purity oxygen to 100 ml / min, and the inlet flow rate of high-purity argon to 900 ml / min. Simultaneously introduce both gases into a 5-liter sealed mixing tank and continue purging for 20 minutes to obtain an oxidizing mixture with an oxygen volume ratio of 10%. This process is a physical gas mixing process and does not involve a chemical reaction. S5. When the furnace temperature reaches 360℃, cut off the forming gas inlet and instantly introduce high-purity argon gas for scavenging for 1.5 minutes. At the same time, adjust the exhaust valve to smoothly transition the total gas pressure in the furnace to 40Pa. After scavenging, cut off the high-purity argon gas inlet within 15 seconds and introduce the oxidizing mixed gas prepared in S4 to accurately establish the oxygen partial pressure in the furnace at 10Pa. Then, maintain the furnace temperature at 360℃ and maintain the oxygen partial pressure at 10Pa for constant temperature heat preservation treatment for 1.5 hours. S6. After the heat preservation is completed, the oxidizing mixed gas inlet is cut off, and high-purity argon gas is introduced into the vacuum annealing furnace until the internal pressure reaches 90 kPa. The heating system is turned off and the single crystal silicon substrate with thin film is cooled to 90°C at a cooling rate of 20°C / min. The furnace is opened and the substrate is taken out to obtain a high-precision thin film resistor.

[0044] Comparative Example 1: Compared with Example 1, the difference is that high-purity nitrogen was not dynamically introduced in S1, and 100% film thickness deposition was completed in a pure high-purity argon atmosphere of 1.5 Pa throughout the process. All other aspects are the same.

[0045] Comparative Example 2: Compared with Example 1, the difference is that the step of "raising to 280°C and holding at a constant temperature under pure vacuum for 3.0 hours" is omitted in S3. After the single crystal silicon substrate is moved into the vacuum annealing furnace and vacuumed, the forming gas is directly introduced and the temperature is raised to 410°C at a heating rate of 5°C / min. All other steps are the same.

[0046] Comparative Example 3: Compared with Example 1, the difference is that the forming gas prepared in S2 is not introduced in S3, but pure high-purity argon gas of the same flow rate is introduced and the absolute pressure is stabilized at 900 Pa before heating. All other aspects are the same.

[0047] Comparative Example 4: Compared with Example 1, the difference is that the step of "introducing high-purity argon gas for 3 minutes" is omitted in S5. After the forming gas is cut off at 410°C, the oxidizing mixed gas prepared in S4 is directly and seamlessly introduced. All other aspects are the same.

[0048] Comparative Example 5: Compared with Example 1, the difference is that after the scavenging in S5, the oxidizing mixed gas prepared in S4 was not introduced, but high-purity argon gas was continued to be introduced to maintain a constant temperature of 410°C and a total pressure of 80Pa for 2.5 hours. All other aspects are the same.

[0049] Test Example 1: Wafer samples from Examples 1 to 5 and Comparative Example 2 that had been sputtered but not yet entered the annealing furnace, as well as blank control substrates from the same batch that were sputtered but had no subsequent heat treatment, were selected. The initial mass of each group of substrates was weighed and recorded using a microbalance. The initial radius of curvature of each wafer surface was measured using a laser curvature scanner. The initial macroscopic residual stress of the thin film was calculated according to the Stoney formula.

[0050] The samples from Examples 1 to 5 were placed in a vacuum annealing furnace for first-stage oxygen-free annealing. After the pure vacuum isothermal holding was completed and before any subsequent gas was introduced, the samples were allowed to cool naturally to room temperature in the furnace before being removed. The samples from each example were weighed again, and the mass change relative to the initial state was calculated. The radius of curvature of the wafer at this stage was measured simultaneously, and the residual stress value in the intermediate state was calculated. Comparative Example 2 lacked the first-stage oxygen-free annealing step, and no sampling test was performed on it at this intermediate stage.

[0051] The sample from the example that completed the intermediate state test was reloaded into the furnace, and the subsequent forming gas pressure holding and heating, micro-oxygen sweeping and passivation, and high-pressure argon quenching steps were performed until the process was completed. Comparative Example 2 was directly executed according to its set direct heating oxidation process. After the finished product was taken out of the furnace, the final wafer curvature of each sample was measured and the final macroscopic residual stress value of the finished product was calculated.

[0052] Table 1: Data on macroscopic residual stress and mass changes of thin films at various stages in the examples and comparative examples Example 1 -512.4 -125.6 -134.2 1.2 Example 2 -489.7 -142.1 -148.5 -0.5 Example 3 -495.2 -131.8 -112.8 0.8 Example 4 -531.1 -119.4 -115.6 2.1 Example 5 -508.6 -138.9 -141.2 0.3 Comparative Example 2 -503.5 - 325.8 - From Table 1, Figure 1 (a) and Figure 1 (b) It is known that all substrates accumulated intrinsic compressive stress in the range of 489.7 MPa to 531.1 MPa during the initial stage of RF deposition. After pure vacuum low-temperature first-stage annealing in Examples 1 to 5, the film compressive stress values ​​decreased to the range of 119.4 MPa to 142.1 MPa. Figure 1 The distribution data in (b) shows that the mass fluctuation of the samples in each embodiment during the first-order oxygen-free annealing stage is between -0.5 μg and 2.1 μg. The constant mass recorded in the experimental test confirms that the pure vacuum heat preservation stage induces microscopic lattice relaxation and vacancy annihilation under a single thermodynamic condition. The absence of an external oxygen source allows the release of physical stress in the film to be independent of the chemical reaction, thus achieving decoupling and separation of stress and chemical oxidation.

[0053] Comparative Example 2, lacking a first-order decoupling stage, directly enters a heated oxidation environment. Figure 1 The stress evolution curve of (a) shows that the internal stress of the finished product reversed from the initial high pressure stress to a tensile stress of 325.8 MPa. The original compressive stress of the film was not released in advance, and the external oxygen invaded the internal grain boundary and chemically combined with the metastable chromium atoms. The volume shrinkage caused by the phase transition and the volume expansion caused by the oxidation reaction produced a superimposed effect, which directly led to selective oxidation of the grain boundary and the formation of chromium-depleted micro-regions inside. After the stress was released in the early stage, the underlying structure of each embodiment tended to be stable. The lattice interstitial steric hindrance effect formed by the in-situ nitriding process was superimposed. The micro-oxygen sweeping passivation reaction triggered by the subsequent introduction of mixed gas was restricted to the surface area of ​​the film. The test results show that the final stress of the finished product of the embodiment was stable at a compressive stress state of 112.8 MPa to 148.5 MPa. This test data confirms the controlled characteristics of the film microstructure and surface chemical state under the intervention of multi-level atmospheres.

[0054] Test Example 2: The wafer samples from Examples 1 to 5, Comparative Examples 3 and 4, which had undergone first-stage oxygen-free annealing and were ready to enter the heating stage, were selected as test objects. Each group of samples was loaded into the vacuum annealing furnace of the quadrupole mass spectrometer through the exhaust pipe. Low heat capacity K-type thin film thermocouples were attached in situ at the geometric center of the surface of each test wafer, and real-time temperature change data of the substrate surface were collected simultaneously.

[0055] During the process section where the equipment is heated to 410°C, a quadrupole mass spectrometer is activated to continuously and dynamically sample the exhaust gas from the annealing furnace. The system continuously extracts and records the peak value of nitrogen emission concentration within this temperature range. During this stage, no forming gas is introduced into the test environment of Comparative Example 3, and only pure high-purity argon is used to maintain the furnace pressure and temperature rise.

[0056] At the moment the process execution reaches the set temperature and the oxidizing mixed gas is introduced, the transient surge value of water vapor concentration obtained by mass spectrometry analysis is recorded. Simultaneously, the maximum span of transient temperature rise on the substrate surface captured by the thin-film thermocouple within a ten-second window before and after gas introduction is extracted. In this test process, the test flow of Comparative Example 4 omits the step of dynamic purging and emptying of high-purity argon gas before the introduction of micro-oxygen mixed gas.

[0057] Table 2: In-situ thermal and mass spectrometry dynamic monitoring data of the examples and comparative examples Example 1 21.3 415.8 12.4 Example 2 18.7 398.2 11.8 Example 3 24.5 432.1 13.2 Example 4 19.8 385.6 10.9 Example 5 22.1 451.3 14.1 Comparative Example 3 1543.6 32.4 1.2 Comparative Example 4 17.5 8952.4 85.6 From Table 2, Figure 2 (a) and Figure 2(b) It can be seen that the monitoring results of the exhaust gas in the heating stage show obvious data differentiation characteristics. In Comparative Example 3, which did not introduce an environment containing hydrogen forming gas, the mass spectrometer probe captured a nitrogen desorption peak of up to 1543.6 ppm. The high temperature environment provides the nitrogen atoms in the interstices of the metal lattice with the kinetic energy to cross the diffusion barrier. A large number of nitrogen atoms migrate to the surface and combine and escape. The purely thermodynamically driven desorption phenomenon directly deconstructs the grain boundary steric barrier formed in the previous process. The corresponding monitoring values ​​of Examples 1 to 5 are all maintained in the low range of 18.7 ppm to 24.5 ppm. The nickel-chromium alloy interface above 200°C triggers the dissociation behavior of hydrogen molecules. The generated chemically adsorbed hydrogen atoms occupy the active sites in the shallow layer and the end of the grain boundary. The hydrogen occupancy at the interface constructs a local kinetic overpressure environment, forming a reverse physical suppression layer, which closes the outward diffusion channels of nitrogen atoms in the internal interstitial space.

[0058] The data on the increase in water vapor and temperature rise at the moment of oxygen introduction further verify the catalytic reaction state at the solid-gas interface. Comparative Example 4 skips the dynamic scavenging step and directly enters the micro-oxygen atmosphere. The free hydrogen gas accumulated in the furnace space and the sudden oxygen molecules trigger a macroscopic gas-phase mixing reaction, resulting in a surge in water vapor concentration to 8952.4 ppm and a transient thermal shock of up to 85.6°C on the wafer surface. The excessively concentrated local heat release causes thermal stress mismatch between the thin film and the substrate, resulting in the loss of self-confined characteristics of the oxidation of surface elements.

[0059] The process in this example involves dynamic purging with high-purity argon gas to remove gaseous free hydrogen, while retaining hydrogen atoms in a limited adsorption and retention state on the low-pressure metal surface. Subsequently, a trace amount of oxygen molecules is introduced to directly react with hydrogen atoms fixed at the grain boundary openings to undergo a heterogeneous catalytic reaction. The increase in water vapor released during this process is controlled within the range of 385.6 ppm to 451.3 ppm, and the induced temperature rise of 10.9°C to 14.1°C is anchored within a few nanometers of the material surface layer. The latent heat of the microscopic reaction released at the interface directly provides the activation energy for the surface chromium elements to cross the nucleation energy barrier, driving the in-situ formation of a high-density passivation layer.

[0060] Test Example 3: Finished wafers from Examples 1 to 5, as well as Comparative Examples 1 and 5, were selected and cut along the cleavage direction using a precision dicing machine to obtain test samples with a size of 10mm × 10mm. Each set of test samples was ultrasonically cleaned to remove dicing debris from the surface, and the surface was dried using a high-purity nitrogen stream. The geometric dimensions of each sample were measured and recorded, and the exposed surface area of ​​each sample was calculated.

[0061] The processed samples were placed in the micro-platinum crucible of the thermogravimetric analyzer. A constant flow of dry pure oxygen at a flow rate of 50 mL / min was continuously introduced into the test chamber. The instrument's programmed temperature module was set to raise the temperature from room temperature to 500℃ at a rate of 10℃ / min, and the temperature was maintained at 500℃ in a pure oxygen environment for 240 minutes.

[0062] Real-time data on the mass change of each sample during the heating and isothermal stages were continuously collected. The absolute weight gain values ​​at the initial node of heating to 500℃, the node of isothermal holding for 120 minutes, and the end node of isothermal holding for 240 minutes were extracted. Combined with the surface area measured before the test, the weight gain value per unit area of ​​the film at each node was calculated.

[0063] Table 3: Thermogravimetric test data under isothermal oxidation environment for the examples and comparative examples Example 1 0.11 0.85 0.91 Example 2 0.09 0.92 0.98 Example 3 0.15 1.09 1.15 Example 4 0.08 0.74 0.80 Example 5 0.12 1.25 1.41 Comparative Example 1 0.55 7.85 11.85 Comparative Example 5 0.25 3.85 5.65 According to Table 3 and Figure 3 The data from the thermogravimetric analyzer, continuously monitored under a pure oxygen environment at 500℃, revealed the evolution of film quality under different process conditions. When the material was in a 500℃ environment, the chemical potential gradient across the gas-solid interface drove oxygen molecules to penetrate deeper into the crystal lattice. The samples from Examples 1 to 5 showed a slight increase in mass in the initial stage of isothermal treatment, and their weight gain curves subsequently slowed down rapidly, converging to 0.80 μg / cm at the end of the test. 2 Up to 1.41 μg / cm 2 Within the range, nitrogen atoms that initially penetrate into the interstices of the thin film lattice construct a local spatial steric hindrance network, which, combined with the micro-oxygen-induced and high-pressure argon-intervention-generated surface dense passivation structure, forms a physical barrier that cuts off the transport channels for oxygen molecules to continuously diffuse into the deep alloy material.

[0064] Compared to other test groups, Comparative Example 1 lacked the in-situ nitriding process, resulting in the internal grain boundaries losing their nitrogen atom occupancy and becoming physically open. During the isothermal period, the weight gain per unit area of ​​this sample continuously increased to 11.85 μg / cm³. 2 The curve shape did not show saturation convergence characteristics. Surface oxygen atoms penetrated and triggered an internal oxidation reaction in the deep matrix alloy. Comparative Example 5 retained the nitrogen steric hindrance effect, but omitted the micro-oxygen mixture passivation step, resulting in the failure to form a dense passivation layer on its surface to resist high-temperature oxidation. The weight gain curve of the sample maintained an upward slope and reached 5.65 μg / cm³ in the later stage of isothermal treatment. 2 The multi-stage atmosphere defect healing treatment implemented on the surface of the thin film, combined with the deep nitrogen atom physical blockage, terminates the continuous oxygen adsorption phenomenon at the interface, thereby controlling the structural degradation phenomenon caused by continuous oxidation of grain boundaries in the alloy thin film under high temperature pure oxygen service conditions from a thermodynamic perspective.

[0065] Test Example 4: Finished wafers from Examples 1 to 5, as well as Comparative Examples 1, 2, and 4, were selected. Using a semiconductor parameter tester with a four-probe stage, the initial sheet resistance of each sample was measured in a constant temperature environment of 25°C. The heating stage was controlled to perform step-up temperature tests within the range of 25°C to 125°C. The resistance-temperature response curves of each sample were extracted, and the initial resistance temperature coefficient was calculated.

[0066] The samples that have completed initial electrical calibration are placed in a high-temperature aging test chamber and subjected to a high-temperature accelerated aging test for a total duration of 1000 hours in an air environment at 150°C. During the test, the samples are removed every 200 hours and allowed to cool in a constant temperature drying cabinet at 25°C for 2 hours. The sheet resistance values ​​at each stage are measured, and the resistance drift data at each aging node are recorded.

[0067] When the aging reaches the end of 1000 hours, the final sheet resistance value of all samples is measured, the cumulative resistance drift rate relative to the initial state is calculated, and a variable temperature test is performed simultaneously in the range of 25℃ to 125℃ to obtain the final initial resistance temperature coefficient after the aging test.

[0068] Table 4: Evolution of Electrical Parameters During High-Temperature Accelerated Aging in Examples and Comparative Cases Example 1 124.5 0.12 15.2 16.8 Example 2 118.7 0.08 12.4 13.5 Example 3 129.3 0.15 18.1 19.3 Example 4 115.6 0.11 14.5 15.2 Example 5 121.8 0.09 13.8 14.9 Comparative Example 1 112.4 2.45 16.5 45.2 Comparative Example 2 135.2 1.85 35.6 68.4 Comparative Example 4 142.1 3.12 28.4 82.1 From Table 4 and Figure 4 (a) and Figure 4 (b) It can be seen that within the 1000-hour test cycle, the final sheet resistance drift rate of Examples 1 to 5 is limited to the low range of 0.08% to 0.15%, and the TCR value changes within 2ppm / ℃ before and after aging. The nitrogen atoms dissolved in the intergranular space during the in-situ nitriding stage reduce the system free energy. Combined with the dense passivation structure generated by multi-stage atmosphere intervention, it forms a physical barrier layer with internal and external bonding. This composite structure blocks the diffusion channel of ambient oxygen to the deep interior of the lattice, inhibits the segregation and migration behavior of metal atoms under the action of thermodynamic field, and maintains the structural stability of the conductive phase.

[0069] The test data of the comparative examples reflect the specific impact of missing process steps on the final electrical stability. In Comparative Example 1, which lacks the in-situ nitriding process, the grain boundaries inside the film are unfilled. Oxygen atoms infiltrate along the grain boundaries under high temperature and undergo internal oxidation. The reduction in the cross-sectional area of ​​the conductive phase leads to a 2.45% resistivity drift, corresponding to a TCR value change from 16.5 ppm / ℃ to 45.2 ppm / ℃. Comparative Example 2, which did not undergo first-order oxygen-free annealing, exhibits higher intrinsic residual stress. The isothermal heating process triggers stress relaxation and... Micrograin boundary slippage results in a 1.85% resistance drift. In Comparative Example 4, which omits the dynamic cleaning operation before micro-oxidation, microcracks in the passivation layer are induced by transient thermal reactions on the surface. The resulting structural defects provide oxygen molecule penetration pathways at 150°C, causing the resistance drift rate to rise to 3.12%. The multi-stage atmosphere intervention process used in the example generates a chemical passivation medium at the solid-gas interface, while eliminating the mechanical internal stress accumulated during the film formation stage. The above-mentioned grain boundary control methods have a definite physical effect in cutting off the high-temperature oxidation path and maintaining the stability of the metal phase structure.

[0070] Test Example 5: Finished wafers from Examples 1 to 5, as well as Comparative Examples 1, 4, and 5, were selected as test subjects. Each wafer was diced into 15mm × 15mm electrochemically polarized samples using a dicing machine. Test leads were soldered to the edges of the samples, and epoxy resin was used to insulate and encapsulate the non-test areas and the edges of the cross-sections, leaving a 1cm core intact. 2 The exposed surface of the thin film.

[0071] The cured encapsulated sample was connected to the three-electrode testing system of the electrochemical workstation. A sodium chloride aqueous solution with a mass fraction of 3.5% was prepared as the corrosion electrolyte. The testing system used a saturated calomel electrode as the reference electrode and a platinum mesh as the auxiliary electrode. The open circuit potential of the sample was measured under a constant temperature of 25°C. The system was allowed to stand and balance for 30 minutes until the potential fluctuation amplitude was less than 2mV.

[0072] 3. Set the electrochemical system to perform potentiodynamic polarization scanning at a scan rate of 1 mV / s within the range of -250 mV to +250 mV relative to the open circuit potential. Extract the polarization response curves of each group of samples. Fit the anodic and cathodic linear intercepts of the polarization curves using the built-in Tafel extrapolation method to calculate the corresponding self-corrosion potential and self-corrosion current density values.

[0073] Table 5: Electrochemical polarization parameter data of the examples and comparative examples in 3.5% NaCl solution Example 1 -123.6 3.2 Example 2 -151.2 4.8 Example 3 -115.4 2.6 Example 4 -134.8 3.9 Example 5 -142.5 4.1 Comparative Example 1 -456.3 156.4 Comparative Example 4 -384.7 89.2 Comparative Example 5 -312.9 45.7 Table 5 and the curves reflecting the electrodynamic polarization Figure 5 and Figure 6It can be seen that the polarization curves of Examples 1 to 5 exhibit typical passivation barrier characteristics, with the self-corrosion potential distributed in the positive high range of -151.2mV to -115.4mV, and the self-corrosion current density maintained at 2.6nA / cm. 2 Up to 4.8 nA / cm 2 At a very low level, the dense passivation layer grown in situ by the multi-stage atmosphere intervention process cuts off the penetration contact of free chloride ions in the solution to the inner surface. Nitrogen atoms that penetrate into the interlattice simultaneously compress the physical depth of the internal grain boundaries. The above-mentioned dual microscopic barrier structure cuts off the liquid phase ion conduction circuit of the local electrochemical galvanic cell and restricts the electron transfer behavior of metal atoms during the anodic dissolution process.

[0074] Under humid electrical stress service conditions, alloy films undergo structural disintegration due to the coupled corrosion of water vapor and free ions. Test data from the comparative group confirms the microscopic physical defects caused by the omission of specific preparation steps. Comparative Example 1, lacking in-situ nitriding treatment, is in an open grain boundary state. Chloride ions with very small ionic radii penetrate deep into the matrix along the grain boundary gaps and induce pitting corrosion. During testing, the anodic branch current of the polarization curve diverges, and the self-corrosion current density reaches 156.4 nA / cm². 2 In Comparative Example 4, the dynamic cleaning operation was omitted, which triggered a transient and intense thermal reaction on the surface. The thermal stress microcracks generated in the passivation layer provided an internal transport channel for the electrolyte, and the corrosion potential shifted negatively to -384.7mV. In Comparative Example 5, the micro-oxygen passivation stage was lacking, and the loose oxide groups remaining on the film surface failed to resist the erosion and replacement effect of halide ions. The grain boundary steric hindrance and surface passivation composite structure used in the examples maintained the chemical inertness of the interface in the polarized electric field, establishing a material physical path to block the deep metal electrochemical corrosion process.

[0075] Test Example 6: Finished wafers from Examples 1 to 5, as well as Comparative Examples 2, 3 and 4, were selected as test objects. Each wafer was cut into 20mm×20mm physical and mechanical test samples using a laser slitting device. The samples were continuously cleaned with ultrasonic acetone and anhydrous ethanol, dried with high-purity nitrogen gas, and placed in a vacuum drying oven for 24 hours to remove surface adsorbed moisture.

[0076] The processed sample was fixed on the five-axis sample stage of a high-resolution X-ray diffractometer. The system used a CuKα ray source and performed wide-angle scanning with a fixed incident angle of 1 degree in grazing incidence mode. The testing process utilized Sin 2 The ψ method involves selecting five different ψ tilt angles within the range of 0 to 60 degrees to collect the diffraction peak displacement data of the (111) crystal plane. Based on the extracted lattice strain rate and the standard elastic constant of the alloy material, the in-plane residual stress value inside the thin film layer of each sample is calculated. The test specifies that the tensile stress is positive and the compressive stress is negative.

[0077] The interfacial bonding strength of the thin film was evaluated using a nano-scratch analyzer equipped with an acoustic emission sensor. A standard Berkovich diamond indenter with a radius of curvature of 2 μm was used to perform a progressive variable load scratch test on the central region of the sample. The normal load of the indenter was set to increase linearly from 0 mN to 80 mN, the scratch loading length was set to 2 mm, and the moving speed was kept constant at 1 mm / min. The system simultaneously collected data on the fluctuation of the friction coefficient and the abrupt change points of the acoustic emission signal during the sliding process, and extracted the critical load value at which the film's interfacial discontinuity peeling failure occurred. The 50 μm × 50 μm area at the end of the scratch was scanned using an atomic force microscope in tapping mode, and the microcrack density per unit area was statistically analyzed.

[0078] Table 6: Test data of residual stress and interfacial mechanical parameters of the thin film in the examples and comparative examples Example 1 -134.2 48.2 0 Example 2 -148.5 51.5 0 Example 3 -112.8 46.8 0 Example 4 -115.6 52.4 0 Example 5 -141.2 49.1 0 Comparative Example 2 +325.8 21.6 15 Comparative Example 3 +142.7 28.5 8 Comparative Example 4 -85.4 18.2 45 Table 6 and the table reflecting the relationship between residual stress and bond strength distribution Figure 7 and Figure 8 It can be seen that the in-plane residual stress of the samples in Examples 1 to 5 is distributed in the compressive stress range of -148.5MPa to -112.8MPa. The peeling critical load extracted by the nano-scratch test is not less than 46.8mN. The multi-stage annealing process excites long-range atomic rearrangement in a vacuum environment, releasing the intrinsic stress generated by the disordered stacking of deposited atoms in the sputtering deposition stage. The controlled occupation of nitrogen atoms in the intergranular gaps of the metal induces micro-lattice expansion in a specific crystal orientation. This intrinsic compressive stress induced by chemical modification offsets the tensile strain field applied by the cooling and shrinkage of the substrate on a macroscopic scale, avoiding the interface debonding behavior triggered by micro-shear forces. Before reaching the critical load failure point, the scratch trajectory edge of the sample in the examples remains smooth and no material debris peeling occurs.

[0079] The comparative test data quantifies the microstructure deterioration caused by the peeling of specific process sections. Comparative Example 2, which omitted the first-stage oxygen-free annealing, retained a high residual tensile stress state of +325.8 MPa. The excessively high intralayer tension weakened the bonding energy of the interface atoms, resulting in a sudden increase in the friction coefficient under a normal load of 21.6 mN and the formation of a continuous peeling zone behind the indenter. Comparative Example 3, due to the lack of forming gas to block desorption, experienced a large number of nitrogen atoms that had previously occupied the grain boundaries escaped due to the high-temperature environment. The local lattice contraction and collapse transformed into a tensile stress field of +142.7 MPa, causing deterioration of the interfacial bonding. Comparative Example 4 lacked dynamic sweeping buffer, resulting in surface vapor phase thermal shock. The rapid thermal expansion and contraction effect induced a density of up to 45 atoms / mm in the passivation layer. 2 The microcrack network, with its deep three-dimensional microcracks disrupting the continuity of stress transmission, reduced the critical load of the sample to 18.2 mN. The aforementioned grain boundary filling and stress reconstruction intervention methods reshaped the micromechanical distribution of the thin film, establishing a structural physical basis with anti-peeling properties.

[0080] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for thin-film vacuum sputtering and heat treatment of a high-precision thin-film resistor, characterized in that, Includes the following steps: S1. Load the single-crystal silicon substrate into the magnetron sputtering chamber for backsputter cleaning; then use the nickel-chromium alloy target as the sputtering source to deposit a thin film in a pure high-purity argon atmosphere. When the deposition thickness reaches the first set thickness of the target total thickness, high-purity nitrogen is dynamically introduced, and sputtering continues in a mixed plasma atmosphere to complete the remaining film thickness deposition, thus obtaining a single-crystal silicon substrate with a thin film attached. S2. High-purity hydrogen and high-purity argon are simultaneously introduced into a sealed mixing tank for physical mixing to obtain a shaped gas. S3. The single-crystal silicon substrate with the thin film is moved into the vacuum annealing furnace, evacuated to the set background vacuum level, heated to the first set temperature, and kept at a constant temperature under pure vacuum. After the heat preservation is completed, forming gas is introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure inside the furnace under the reducing pressure holding state of the set pressure and continue to heat up. S4. High-purity oxygen and high-purity argon are simultaneously introduced into a closed mixing tank for physical mixing to obtain an oxidizing mixture. S5. When the furnace temperature reaches the second set temperature, cut off the forming gas inlet and introduce high-purity argon gas for purging. At the same time, adjust the exhaust valve to reduce the total gas pressure in the furnace. After purging, cut off the high-purity argon gas inlet and introduce oxidizing mixed gas to establish the oxygen partial pressure in the furnace at the set partial pressure value and perform constant temperature heat preservation treatment. S6. After the heat preservation is completed, the oxidizing mixed gas inlet is cut off, and high-purity argon gas is introduced into the vacuum annealing furnace for pressurization and cooling. The furnace is then opened and the product is removed to obtain a high-precision thin-film resistor.

2. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S1, the cleaning process includes: loading a single-crystal silicon substrate with a thermally oxidized silicon insulating layer on its surface into a magnetron sputtering chamber, evacuating the base vacuum of the magnetron sputtering chamber to a set sputtering base vacuum, introducing high-purity argon gas to maintain the absolute pressure in the chamber at 0.5 Pa to 1.5 Pa, and starting the radio frequency power supply to perform reverse sputtering cleaning on the single-crystal silicon substrate for 5 to 15 minutes.

3. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S1, the in-situ nitriding deposition process includes the following steps: After the cleaning process, a nickel-chromium alloy target composed of 40 to 80 parts by weight of nickel and 20 to 60 parts by weight of chromium is used as a sputtering source. A thin film is deposited in a pure high-purity argon atmosphere with a sputtering power density of 2.0 W / cm2 to 8.0 W / cm2 and a sputtering pressure of 0.5 Pa to 1.5 Pa. When the deposition thickness reaches 85% to 95% of the target total thickness, high-purity nitrogen is dynamically introduced so that the volumetric flow rate accounts for 3% to 8% of the total inlet flow rate. Sputtering continues in a mixed plasma atmosphere to complete the deposition of the remaining 5% to 15% of the film thickness, resulting in a single-crystal silicon substrate with a thin film.

4. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S2, the gas mixing process includes: under the conditions of setting the mixing temperature and setting the mixing pressure, turning on the gas mass flow meter, setting the inlet flow rate of high-purity hydrogen to 30 ml / min to 100 ml / min, setting the inlet flow rate of high-purity argon to 900 ml / min to 970 ml / min, and simultaneously introducing the above two gases into a sealed mixing tank with a set volume, continuously introducing gas for a set mixing time to obtain a forming gas with a hydrogen volume ratio of 3% to 10%. This process is a physical gas mixing process and no chemical reaction occurs.

5. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S3, the first-order oxygen-free annealing step includes: transferring the single-crystal silicon substrate with the thin film into a vacuum annealing furnace, evacuating the furnace to the set annealing vacuum level, raising the temperature to 220°C to 280°C at a heating rate of 3°C / min to 8°C / min, and holding it at a constant temperature under pure vacuum for 1.5 hours to 3.0 hours.

6. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S3, the subsequent forming gas pressure holding and heating steps include: after the heat holding is completed, the furnace temperature is kept constant, and the forming gas prepared in step S2 is introduced into the vacuum annealing furnace to establish and stabilize the absolute pressure in the furnace at 400 Pa to 900 Pa. Under this reducing pressure holding state, the furnace temperature is raised from 220°C to 280°C to 350°C to 410°C at a heating rate of 2°C / min to 5°C / min.

7. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S4, the gas mixing process includes: under the conditions of setting the mixing temperature and setting the mixing pressure, turning on the gas mass flow meter, setting the inlet flow rate of high-purity oxygen to 50 ml / min to 200 ml / min, setting the inlet flow rate of high-purity argon to 800 ml / min to 950 ml / min, and simultaneously introducing the above two gases into a sealed mixing tank with a set volume, continuously introducing gas for a set mixing time to obtain an oxidizing mixture with an oxygen volume ratio of 5% to 20%. This process is a physical gas mixing process and no chemical reaction occurs.

8. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S5, the gas phase buffer purging step includes: when the furnace temperature reaches 350℃ to 410℃, the forming gas inlet is cut off, and high-purity argon gas is introduced instantaneously for purging for 1 to 3 minutes, while adjusting the exhaust valve to reduce the total gas pressure in the furnace to 20Pa to 80Pa.

9. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S5, the subsequent isothermal oxidation passivation steps include: after the cleaning is completed, the high-purity argon gas inlet is cut off within 10 to 30 seconds, and the oxidizing mixed gas prepared in step S4 is introduced to accurately establish the oxygen partial pressure in the furnace at 8 Pa to 20 Pa. Then, the furnace temperature is maintained at 350°C to 410°C and the oxygen partial pressure is maintained at 8 Pa to 20 Pa for isothermal holding treatment for 1.0 hour to 2.5 hours.

10. The thin-film vacuum sputtering and heat treatment method for high-precision thin-film resistors according to claim 1, characterized in that, In step S6, the pressurization and cooling steps include: after the heat preservation is completed, the oxidizing mixed gas inlet is cut off, high-purity argon gas is introduced into the vacuum annealing furnace until the internal pressure reaches 80 kPa to 110 kPa, the heating system is turned off and the single crystal silicon substrate with thin film is cooled to the set furnace exit temperature at a cooling rate of 15℃ / min to 30℃ / min, the furnace is opened and the substrate is taken out to obtain a high-precision thin film resistor.

Images