Method of designing, method of manufacturing and cleaning end point detection device for detection chamber

Abstract

The application relates to the technical field of semiconductor detection, and discloses a design method and a preparation method of a detection cavity and a cleaning end point detection device, which comprises the following steps: acquiring the molecular structure of a to-be-detected gas, a working temperature, and the surface energy of each candidate inner wall material under a plurality of different surface terminals in a plurality of candidate inner wall materials; simulating and determining the adsorption energy, the desorption activation energy and the desorption parameters between the molecular structure of the to-be-detected gas and the surface terminals corresponding to the plurality of surface energies under the working temperature; selecting a target inner wall material from the plurality of candidate inner wall materials and selecting a target surface terminal from the plurality of surface terminals, and the target adsorption energy is within-0.5 eV to-0.1 eV; determining the target desorption temperature of the detection cavity in the cleaning process stage based on the target desorption activation energy; and storing the target desorption parameters corresponding to the target inner wall material and the target surface terminal. The application can inhibit the memory effect, inhibit the baseline drift and shorten the response time.

Description

Technical Field

[0001] This invention relates to the field of semiconductor testing technology, and specifically to a method for designing and fabricating a detection cavity and a device for detecting the cleaning endpoint. Background Technology

[0002] Non-dispersive infrared (NDIR) cleaning endpoint detection devices include a detection cavity for carrying the optical path and gas channel. As the critical dimensions of semiconductors continue to shrink, especially after entering deep submicron and nanometer process nodes, the interfacial adsorption of gas molecules between the cavity inner wall materials remains significant, which is even more detrimental to advanced process detection. This interfacial adsorption has the following three problems in semiconductor cleaning endpoint detection devices: The first aspect is the memory effect. Gas molecules from a previous measurement remain on the inner wall of the chamber and slowly desorb during subsequent measurements, superimposing themselves on the current concentration signal and causing the reading to be too high. In the integrated circuit manufacturing process, the memory effect caused by trace gaseous contaminants adsorbed on the sample tubing or inner wall of the chamber directly leads to a severe decrease in measurement accuracy, making it difficult to meet the requirements for real-time monitoring of contaminants at the part per billion (PPB) / part per trillion (PPT) level in semiconductor cleanrooms.

[0003] Secondly, the response time is prolonged. The adsorption / desorption kinetic imbalance causes the settling time of the cleaning endpoint detection device when switching from one gas concentration to another to be much longer than the mechanical ventilation time of the chamber. In in-situ monitoring of semiconductor process equipment, such as online analysis of etching chamber byproducts, it is difficult to meet the requirements for high-time-resolution rapid sequence measurements.

[0004] The third aspect involves baseline drift and linearity degradation. Temperature fluctuations caused by the cyclic heating and cooling of process chambers in wafer fabs alter the adsorption / desorption balance, leading to zero-point drift. More importantly, the adsorption process itself may change the surface properties of the inner wall (such as roughness and reflectivity) or release trace amounts of interfering substances, causing baseline drift, signal attenuation, or increased background noise in the non-dispersive infrared cleaning endpoint detection device, thereby reducing the sensitivity and accuracy of detecting low-concentration gases. Long-term use and contaminant accumulation exacerbate the nonlinearity of the sensor response curve, making it difficult to meet the stability requirements for continuous operation for months.

[0005] In related technologies, high-temperature cleaning or inert coating methods are used to overcome interfacial adsorption and avoid the aforementioned problems. High-temperature cleaning involves heating the detection cavity between measurements to force desorption. The heating temperature is 150℃ to 200℃ or even higher, lasting from several seconds to several minutes. However, high-temperature treatment not only consumes a large amount of energy and reduces the thermal cycle life of the device, but excessively high temperatures may also damage the thin-film structure on non-dispersive infrared cleaning endpoint detection devices, such as the micro-hot plate of the infrared light source, the sensitive layer of the pyroelectric detector, etc., and even cause a decrease in the reliability of solder joints or bonding interfaces in wafer-level packaging.

[0006] Inert coating methods involve sputtering and depositing an inert metal layer, such as gold (Au) or platinum (Pt), onto the inner wall of the cavity, or coating the inner wall with polytetrafluoroethylene (PTFE) to suppress adsorption by reducing surface activity. However, for polar gases such as carbon dioxide (CO2) and nitrogen dioxide (NO2), the cumulative effect of physical adsorption still leads to a significant memory effect. While sputtering a metal reflective layer onto the inner wall of the silicon cavity improves light reflection efficiency to some extent, it does not fundamentally solve the problem of gas molecule adsorption on the metal surface. Furthermore, the line-of-sight limitation of the sputtering process makes it difficult to ensure coating uniformity on the inner wall of deep and narrow structures. Summary of the Invention

[0007] This invention provides a method for designing and fabricating a detection cavity, as well as a cleaning endpoint detection device, to solve the problems of memory effect, prolonged response time, baseline drift, and linearity degradation caused by adsorption without affecting device performance.

[0008] In a first aspect, the present invention provides a method for designing a detection cavity for introducing a test gas to detect the cleanliness endpoint of a semiconductor process chamber. The design method includes: acquiring the molecular structure, operating temperature, and surface energy of each candidate inner wall material at multiple different surface terminals among multiple candidate inner wall materials, wherein the surface terminal is the atomic layer in contact with the test gas after the inner wall of the detection cavity is functionalized by an atomic layer deposition process; at the operating temperature, simulating and determining the adsorption energy, desorption activation energy, and desorption parameters between the molecular structure of the test gas and the surface terminals corresponding to the multiple surface energies, wherein the desorption parameters include the diffusion coefficient and the near-wall desorption time determined by the diffusion coefficient; selecting a target inner wall material from multiple candidate inner wall materials, and selecting a target surface terminal from multiple surface terminals, wherein the target inner wall material acts on the target surface terminal. The adsorption energy between the gas and the target gas molecular structure meets the first condition, and the theoretical desorption time determined by the desorption activation energy meets the second condition. The adsorption energy meeting the first condition is taken as the target adsorption energy, and the desorption activation energy meeting the second condition is taken as the target desorption activation energy. The first condition is that the target adsorption energy is within -0.5 eV to -0.1 eV, and the second condition is that the theoretical desorption time is within 1 s to 5 s. Based on the target desorption activation energy, the target desorption temperature of the detection chamber during the cleaning process is determined. Desorption parameters corresponding to the target inner wall material and the target surface terminal are stored as target desorption parameters to obtain the design parameters corresponding to the detection chamber. The target desorption parameters are used to determine the signal deviation caused by gas molecules remaining on the inner wall during the detection stage. The design parameters include the target inner wall material, the target surface terminal, the target desorption temperature, and the target desorption parameters.

[0009] The adsorption energy determines the desorption activation energy, which in turn determines the desorption rate constant via the Arrhenius equation. This rate constant, combined with the diffusion coefficient, determines the near-wall desorption time. In a continuous flow system, the near-wall desorption time is amplified by the adsorption retention factor to ultimately determine the theoretical desorption time. The theoretical desorption time is a physical prediction calculated using the Arrhenius equation, and its range (1 s to 5 s) serves as the target desorption time for actual parameter control of the detection chamber. The theoretical desorption time characterizes the theoretical value of the expected completion time of the desorption process calculated based on the adsorption-desorption kinetic model (including physical parameters such as adsorption energy, temperature, and diffusion coefficient), reflecting the mathematical prediction results of the physical model. The target desorption time, on the other hand, characterizes the preset benchmark value used by the controller to compare with the detected actual desorption time in actual process control, reflecting the process control objective that the control system aims to achieve. The former emphasizes the source of theoretical calculations, while the latter emphasizes actual control. Only when the adsorption energy, desorption activation energy, and near-wall desorption time all meet their respective limiting conditions can the memory effect of the detection cavity be effectively eliminated, and the detection results of the cleaning endpoint be accurate and reliable.

[0010] In one alternative embodiment, the surface terminal is a hydrogen terminal, a hydroxyl terminal, an oxygen terminal, a chlorine terminal, an alkyl terminal, or a fluoroalkyl terminal.

[0011] In one alternative embodiment, the candidate inner wall material is silicon nitride, yttrium oxide, silicon carbide, stainless steel, alumina, or polytetrafluoroethylene.

[0012] In one optional embodiment, the near-wall concentration boundary layer thickness of the detection chamber ranges from 6.25 μm to 10 μm. The near-wall concentration boundary layer is a fluid layer that adheres to the inner wall surface of the detection chamber and has a concentration gradient. The near-wall concentration boundary layer thickness is used to characterize the equivalent distance of the near-wall mass transfer resistance concentration region. The near-wall concentration boundary layer thickness is determined by the following formula: δ=l / Sh; Where δ represents the thickness of the near-wall concentration boundary layer, l is the equivalent feature size of the detection cavity, Sh is the Sherwood number, and the equivalent feature size is the cross-sectional feature size of the detection cavity.

[0013] The equivalent characteristic size determines the mass transfer boundary layer development of the detection cavity. The value of the equivalent characteristic size l of the detection cavity ranges from 25 mm to 40 mm, and the Sherwood number Sh is 4000, which is determined based on the turbulent mass transfer near-wall region model.

[0014] In one optional implementation, determining the target desorption temperature of the detection chamber during the cleaning process based on the target desorption activation energy includes: determining the target desorption temperature using the following formula: ; In the formula, Indicates the target desorption temperature. Indicates the target desorption activation energy. Represents the gas constant. , This indicates the frequency of attempts by surface-adsorbed molecules. This indicates the target desorption time, and the target desorption temperature is the thermal control threshold of the cleaning process stage, in order to complete the surface molecule desorption within the target desorption time.

[0015] In one alternative implementation, the diffusion coefficient is greater than or equal to Alternatively, the near-wall desorption time is less than or equal to 50 ms.

[0016] Secondly, the present invention provides a method for fabricating a detection cavity, the method comprising: providing a semiconductor structure; forming an inner wall layer with a thickness of 10 μm to 100 μm and made of a target inner wall material on the inner surface of the semiconductor structure by a thin film growth process; wherein the semiconductor structure includes a columnar body portion having a receiving cavity, a first opening and a second opening, the first opening and the second opening being opposite to and communicating with the receiving cavity; the sidewall of the body portion also having a first air inlet and a first air outlet communicating with the receiving cavity; placing the semiconductor structure with the grown inner wall layer on a carrier plate of a process chamber; wherein, in the current N deposition cycles, the second air inlet of the process chamber is communicating with the first opening, and the second air outlet of the process chamber is communicating with the second opening, so as to form an axial airflow path in the receiving cavity; in the next set of N deposition cycles... Inside, the second air inlet is connected to the second opening, and the second air outlet is connected to the first opening, where N is an integer. Using atomic layer deposition (ALD), precursors and reactants are alternately introduced into the process chamber according to the target pulse time and target purge time, depositing a functionalized coating on the inner wall layer to obtain an initial surface terminal. The target pulse time and target purge time are determined based on the length of the semiconductor structure and the diffusion coefficient of the precursor. Water vapor is introduced into the process chamber to heat-insulate the semiconductor structure with the initial surface terminal, performing in-situ surface conditioning to ensure the conditioned initial surface terminal meets the first condition, thus obtaining the target surface terminal. Nitrogen gas is introduced and annealing is performed to prepare the detection chamber. The design parameters corresponding to the detection chamber are determined by the detection chamber design method of the first aspect or any corresponding embodiment described above.

[0017] In one alternative embodiment, the material of the functionalized coating includes at least one of titanium dioxide, aluminum oxide, and zinc oxide, and the thickness of the functionalized coating is from 5 nm to 20 nm.

[0018] In one alternative embodiment, the substrate temperature of the atomic layer deposition process is 50°C to 300°C, and the background vacuum is less than or equal to 1×10⁻⁶. -3 The operating pressure is 0.1 Torr to 5 Torr, the target purging time is 5s to 60s, the total number of deposition cycles is 50 to 200, and the value of N is 5 to 20.

[0019] In one alternative embodiment, the pressure of the steam is 0.1 Torr to 1 Torr, and the heat treatment duration is 10 min to 30 min.

[0020] Thirdly, the present invention provides a clean endpoint detection device, which includes an infrared light source, an infrared detector, a controller, and a detection cavity. The detection cavity is prepared by the detection cavity preparation method described in the second aspect or any corresponding embodiment above. The infrared light source is used to generate a first optical signal of a target wavelength and a second optical signal of a reference wavelength. The gas to be tested is introduced into the detection cavity. The first and second optical signals are transmitted to the infrared detector after passing through the detection cavity. The first optical signal is absorbed to different degrees according to the concentration change of the gas to be tested. The infrared detector is used to convert the received first optical signal into a first electrical signal and the received second optical signal into a reference electrical signal. The controller is used to determine the absorption rate based on the first electrical signal and the reference electrical signal to obtain the concentration of the gas to be tested.

[0021] In one optional embodiment, the cleaning endpoint detection device further includes a temperature sensing module, and the controller stores target desorption parameters. The controller is also used to obtain the current temperature of the detection chamber from the temperature sensing module and to obtain the measured concentration of the gas to be tested in the detection chamber. A kinetic model is constructed based on the current temperature, measured concentration, and target desorption parameters. The signal deviation caused by the retention of gas molecules on the inner wall of the detection chamber is determined through the kinetic model. The deviation of the measured concentration of the gas to be tested is corrected based on the signal deviation to obtain the corrected concentration.

[0022] In one optional embodiment, the cleaning endpoint detection device further includes a heating module. The controller is configured to: before detection, control the heating module to heat the detection chamber to desorb gas molecules adsorbed on the inner wall of the detection chamber, thereby eliminating the memory effect; during the heating and desorption process, acquire the detection signal of the detection chamber; when the detection signal recovers from the stable concentration value along the falling edge to the clean state baseline value, extract the time interval of the falling edge of the detection signal as the actual desorption time; and adjust the current heating temperature of the heating module according to the deviation between the actual desorption time and the target desorption time; wherein, the target desorption time is the total equivalent time from the start of heating of the detection chamber to the wall adsorption, molecule desorption and removal by the vacuum pump, and the recovery of the detection signal to the clean state baseline value, including the near-wall molecule desorption time and the continuous gas intake gas replacement time, and the gas replacement time is represented by the average extraction time of a single tube.

[0023] Near-wall desorption time is the characteristic time required for gas molecules to move from their adsorption state on the inner wall of the detection chamber, desorb, and cross the near-wall concentration boundary layer before entering the bulk gas phase. The average single-tube extraction time is the average time required for gas molecules to be extracted from the detection chamber by the vacuum pump after entering the bulk gas phase; its value is determined by the ratio of the detection chamber volume to the effective pumping speed of the vacuum pump. Under continuous flow conditions with continuous gas intake, the macroscopic extraction process consists of multiple microscopic extraction events of gas from individual tubes. The target desorption time is the equivalent magnified value of the sum of the near-wall desorption time and the average single-tube extraction time.

[0024] In one alternative implementation, the controller determines the deviation ratio using the following formula: p = (t) actual -t target ) / t target ; Where p represents the deviation ratio, t actual t represents the actual desorption time of gas molecules within the detection chamber. target This indicates the target desorption time; the controller adjusts the current heating temperature of the heating module according to the deviation ratio; where, when t actual >t target When p > 0, the controller will increase the current heating temperature by K × p × T. step The adjusted current heating temperature is: T adjust =T current +K×p×T step When t actual <t target When p < 0, the controller will reduce the current heating temperature by K × |p| × T. step The adjusted current heating temperature is: T adjust =T current -K×|p|×T step When t actual =t target When p=0, the controller maintains the current heating temperature unchanged; T current T represents the current heating temperature of the heating module. adjust T represents the adjusted current heating temperature. step This indicates the preset temperature step size, and K is the gain coefficient. The gain coefficient is a preset constant or determined by the controller based on the deviation ratio. The value of K ranges from 1 to 8.

[0025] Fourthly, the present invention provides a gas concentration measurement method for a cleaning endpoint detection device, wherein the cleaning endpoint detection device is the cleaning endpoint detection device described in the third aspect above or any corresponding embodiment thereof, and the gas concentration measurement method is applied to the controller of the cleaning endpoint detection device, and the gas concentration measurement method includes: Heating is controlled by the heating module at the current heating temperature to desorb residual gas molecules on the inner wall of the detection chamber. After the current temperature of the detection chamber stabilizes, nitrogen gas is introduced into the detection chamber, and the infrared detector detects the reference electrical signal and performs zero-point calibration. The gas to be tested is introduced into the detection chamber, and the infrared light source is controlled to generate a first optical signal of the target wavelength and a second optical signal of the reference wavelength. The first and second optical signals are transmitted to the infrared detector after passing through the detection chamber. The gas to be tested absorbs the first optical signal, and the infrared detector converts the received first optical signal into a first electrical signal and the received second optical signal into a reference electrical signal. The absorption rate is determined based on the first and reference electrical signals to obtain the concentration of the gas to be tested. The current temperature of the detection chamber and the measured concentration of the gas to be tested in the detection chamber are obtained. Based on the kinetic model constructed using the current temperature, measured concentration, and target desorption parameters, the signal deviation caused by the retention of gas molecules on the inner wall of the chamber is determined. The deviation of the measured concentration of the gas to be tested is corrected based on the signal deviation to obtain the concentration of the gas to be tested.

[0026] Fifthly, the present invention provides an electronic device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the design method of the detection cavity of the first aspect or any corresponding embodiment described above, or to perform the gas concentration measurement method of the cleaning endpoint detection device of the fourth aspect or any corresponding embodiment described above.

[0027] In a sixth aspect, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the design method of the detection cavity of the first aspect or any corresponding embodiment thereof, or to execute the gas concentration measurement method of the cleaning endpoint detection device of the fourth aspect or any corresponding embodiment thereof.

[0028] In a seventh aspect, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to execute the design method of the detection cavity of the first aspect or any corresponding embodiment thereof, or to execute the gas concentration measurement method of the cleaning endpoint detection device of the fourth aspect or any corresponding embodiment thereof.

[0029] The design method, preparation method, and cleaning endpoint detection device for the detection cavity provided by this invention have at least the following advantages: This invention uses multi-scale simulation to screen for adsorption energies. The first condition is met ( and the time for theoretical decoupling The second condition is met ( The target inner wall material and target surface terminal are modified to modify the inner wall of the detection cavity, so that the interaction between gas molecules and the inner wall surface is dominated by reversible physical adsorption, which suppresses the memory effect and shortens the response time at the source.

[0030] After determining the target inner wall material and the target surface terminal, the target desorption temperature is determined by the target desorption activation energy corresponding to the target inner wall material and the target surface terminal. This ensures that the gas molecules adsorbed on the surface can be completely desorbed within 1 to 5 seconds, and also avoids thermal stress damage to the complementary metal-oxide-semiconductor (CMOS) circuits, micro-electro-mechanical system (MEMS) thin film structures and bonding interfaces on the wafer-level integrated chip caused by empirical high-temperature heating.

[0031] By storing target desorption parameters corresponding to the target inner wall material and the target surface terminal for dynamic signal compensation, residual adsorption interference can be eliminated to the maximum extent during detection, thereby achieving active suppression of baseline drift during long-term operation and meeting the stability requirements of semiconductor process equipment for continuous operation for several months.

[0032] In the scenario of detecting trace gases at the part-in-a-billion level, which is common in advanced semiconductor processes, the first condition is narrowed down to... It can meet the signal stability requirements at lower concentrations.

[0033] Atomic layer deposition (ALD) is used to form a functionalized coating on the inner wall of a semiconductor structure. By utilizing a self-limiting surface reaction mechanism, the coating thickness and surface terminal density can be precisely controlled at the nanoscale. By combining size-matched pulse and purge times with alternating precursor introduction directions between deposition cycles, a uniform, conformally covered atomic-level functional layer can be achieved on the macroscopic aspect ratio cavity inner wall. This effectively suppresses the adsorption and retention effect of gas molecules on the cavity inner wall, reduces the memory effect of physical adsorption, and consequently lowers baseline drift of the detection signal, ensuring long-term stability and repeatability in semiconductor chamber cleaning detection scenarios under low concentration conditions.

[0034] In this invention, by controlling the thickness of the functionalized coating to 5 nm or more during deposition, the continuity of the functionalized coating can be ensured, defects such as pinholes or dislocations can be avoided, and the memory effect can be effectively suppressed. By controlling the thickness of the functionalized coating to 20 nm or less, the integrity of the coating can be avoided due to microcracks caused by high internal stress, and the deposition time can also be avoided.

[0035] This invention limits the substrate temperature to within 50°C to 300°C, allowing for precise control of the proportion and density of chemical species on the surface terminals and the crystallization state of the thin film, thereby locking the target adsorption energy at [temperature range]. to Within the middle range.

[0036] During the long-term operation of the cleaning endpoint detection device, the baseline recovery time is used as a quantitative indicator to sense the coating contamination and aging degree in real time. The current heating temperature and / or the current cleaning time are adjusted according to the baseline recovery time. The device can autonomously correct baseline drift and maintain the detection linearity of multiple concentration points. This ensures that even when the desorption kinetics of the coating surface change due to contamination or oxidation during long-term operation of the sensor, the device can still maintain optimal cleaning efficiency and low power consumption characteristics through parameter self-adjustment. Attached Figure Description

[0037] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0038] Figure 1 This is a flowchart illustrating a method for designing a detection cavity according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a detection cavity according to an embodiment of the present invention; Figure 3 This is a schematic flowchart of a method for preparing a detection cavity according to an embodiment of the present invention; Figure 4 This is a schematic diagram of a cleaning endpoint detection device according to an embodiment of the present invention; Figure 5 This is a schematic diagram of another cleaning endpoint detection device according to an embodiment of the present invention; Figure 6 This is a schematic flowchart of a gas concentration measurement method using a cleaning endpoint detection device according to an embodiment of the present invention. Figure 7 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0041] This invention provides a method for designing and preparing a detection cavity, as well as a cleanliness endpoint detection device. In the design stage of the detection cavity, starting from the underlying physicochemical interactions, the selection of the cavity inner wall material and surface terminals is used as design variables. Adsorption behavior is predicted through multi-scale quantitative simulation. The adsorption energy between the target inner wall material and the molecular structure of the analyte gas under the action of the target surface terminals meets a first condition, and the theoretical desorption time determined by the desorption activation energy meets a second condition. The adsorption energy meeting the first condition is taken as the target adsorption energy, and the desorption activation energy meeting the second condition is taken as the target desorption activation energy. This ensures that the adsorption energy... The first condition is met ( ) and theoretical desorption time The second condition is met ( Using these criteria, the target inner wall material and target surface terminal are selected to achieve a reversible physical adsorption-dominated interaction between the inner wall of the cavity and gas molecules. This suppresses the memory effect, shortens the response time, and reduces cleaning power consumption, meeting the stringent requirements of advanced semiconductor processes for high-precision, rapid response, and long-term stable monitoring of trace gases.

[0042] The detection chamber in this invention is used to introduce the gas to be tested in order to detect the cleaning endpoint of the semiconductor process chamber. The concentration change of the gas to be tested is detected using the non-dispersive infrared principle to provide feedback on the cleaning endpoint. The non-dispersive infrared cleaning endpoint detection device utilizes the selective absorption characteristics of different gas molecules to infrared radiation of specific wavelengths to quantitatively determine the gas concentration based on Lambert-Beer's law. The mathematical expression of Lambert-Beer's law can be shown in formula (1):

[0043] In the formula, Intensity of incident infrared light The intensity of transmitted light. The absorption coefficient of the gas is denoted as . For gas concentration, This is the optical path length.

[0044] The design method of the detection cavity provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0045] Figure 1 This is a flowchart illustrating a method for designing a detection cavity according to an embodiment of the present invention, which can be used in electronic devices, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain the molecular structure, operating temperature, and surface energy of each candidate inner wall material at multiple different surface terminals among multiple candidate inner wall materials.

[0046] For example, the molecular structure of the gas to be tested can be the molecular structure of gases such as carbon dioxide (CO2), nitrogen dioxide (NO2), water (H2O), and ammonia (NH3), and the candidate inner wall material can be silicon nitride, yttrium oxide, silicon carbide, stainless steel, alumina, or polytetrafluoroethylene. Among them, stainless steel can be 316L stainless steel.

[0047] The surface terminal can be a hydrogen terminal (-H terminal), a hydroxyl terminal (-OH terminal), an oxygen terminal (-O terminal), a chlorine terminal (-Cl terminal), an alkyl terminal (-CH3 terminal), or a fluoroalkyl terminal (-CF3 terminal).

[0048] The surface terminal is the atomic layer that comes into contact with the gas to be measured after the inner wall of the detection chamber has been functionalized using atomic layer deposition (ALD) technology. In this invention, after the inner wall material is processed, the atoms on its surface form dangling bonds, which are highly unstable. To stabilize the surface and impart specific properties, different atoms or groups are introduced to combine with the dangling bonds, thereby forming different surface terminals. The outermost atomic layer that comes into contact with the gas after the functionalized coating (such as a transition metal oxide) is deposited is the surface terminal.

[0049] This invention incorporates the chemical structure of the surface terminal as an independent design variable into the screening system. By comparing the interaction energy between the surface of the material modified with different surface terminals under different candidate inner walls and gas molecules, the geometric configuration and electronic structure of the adsorption sites of gas molecules on different surface terminals are quantitatively determined.

[0050] Operating temperature refers to the temperature of the detection chamber when detecting the concentration of the gas to be measured. The adsorption energy determined by subsequent simulations is based on the same candidate inner wall material with different crystal orientations and structures bonded to a specific surface terminal. They may differ, surface energy It is a comprehensive thermodynamic property exhibited by all the microscopic atomic arrangement states. Surface energy can lock a unique interface, thereby accurately calculating the parameters that characterize the adsorption behavior of surface terminals and gas molecules.

[0051] The operating temperature, the molecular structure of the gas to be tested, and the surface energy of each candidate inner wall material at multiple different surface terminals can be selected and set according to the actual process application. The operating temperature, the molecular structure of the gas to be tested, and the multiple surface energies can be input by the engineer into the electronic device through the human-machine interface, so that the electronic device can obtain the operating temperature, the molecular structure of the gas to be tested, and the multiple surface energies.

[0052] It should be understood that the adsorption behavior of the molecular structure of the gas to be tested on the inner wall surface of the detection chamber essentially depends on the matching relationship between the electronic structure (polarity, polarizability, electron affinity) of the gas molecules and the chemical properties of the surface terminals. The design method of this invention is precisely aimed at this matching relationship, and by screening different surface terminals, the adsorption energy can be precisely controlled to a physical adsorption window of -0.5 eV to -0.1 eV.

[0053] Highly polar / highly reactive gas molecules such as nitrogen dioxide, sulfur dioxide (SO2), and ammonia have high electron affinity or lone pairs of electrons, readily undergoing electron transfer to surfaces and forming strong chemisorption (<-0.5 eV). This excessive adsorption leads to difficult desorption and a significant memory effect. Low surface energy inert terminals should be selected to shield the chemically active sites and weaken the excessive adsorption. Fluoroalkyl terminals, with their extremely strong hydrophobic and oleophobic properties, can form a physical isolation layer, effectively inhibiting electron transfer; alkyl terminals can reduce surface energy while retaining a certain degree of physical adsorption capacity.

[0054] Weakly polar / nonpolar gas molecules such as carbon dioxide, methane (CH4), and carbon monoxide (CO) have stable molecular structures and interact weakly with surfaces through van der Waals forces. The adsorption energies on bare surfaces are often below -0.1 eV, resulting in insufficient signal response. Therefore, it is necessary to increase the adsorption energies to detectable levels. Polar terminals should be selected to provide suitable active sites, enhancing physical adsorption through dipole-induced dipole interactions. Hydroxyl terminals provide uniform polar sites and have mature preparation processes; oxygen terminals can provide a stronger polar environment, but density control is necessary to avoid excessive adsorption.

[0055] In specific process environments with high humidity or multiple interfering gases, competitive adsorption of water molecules often severely interferes with gas detection. Hydrophobic terminals are chosen to repel polar water molecules, while a physical barrier layer reduces non-specific adsorption. Fluoroalkyl terminals are suitable for this scenario; their dual hydrophobic and oleophobic properties effectively resist moisture and contamination. Alkyl terminals also form hydrophobic surfaces.

[0056] Step S102: At the operating temperature, simulate and determine the adsorption energy, desorption activation energy, and desorption parameters between the molecular structure of the gas to be tested and the surface terminals corresponding to multiple surface energies.

[0057] The desorption parameters include the diffusion coefficient D and the near-wall desorption time. Alternatively, the desorption parameter could be the near-wall desorption time. .

[0058] This invention utilizes density functional theory (DFT) within the Vienna Ab-initio Simulation Package (VASP) to simulate and determine the adsorption energies between the molecular structure of the gas under test and multiple surface terminals corresponding to different surface energies at the operating temperature. .

[0059] This invention can use molecular dynamics (MD) theory to simulate and determine the desorption activation energy between the molecular structure of the gas being tested and multiple surface terminals corresponding to different surface energies in a large-scale atomic / molecular massively parallel simulator (LAMMPS). The adsorption and desorption parameters can also be obtained by simulating the adsorption energy using the Reactive Force Field (ReaxFF) molecular dynamics theory. Desorption activation energy And desorption parameters.

[0060] The electronic device is equipped with simulation software (Vienna First Principles Simulation Software Package, Large-Scale Atom / Molecular Massive Parallel Simulator). The input parameters of the simulation software include the molecular structure of the gas to be tested, the surface energy between each candidate inner wall material and each surface terminal in the multiple candidate inner wall materials to be simulated, and the operating temperature.

[0061] Molecular dynamics can be reactive force field molecular dynamics, condensed-phase optimized molecular potentials for atomistic simulation studies (COMPASS) molecular dynamics, or optimized potentials for liquid simulations all atom (OPLS-AA) molecular dynamics.

[0062] The reaction force field is an advanced interatomic potential energy model. Unlike traditional force fields, the core advantage of the reaction force field is that it can simulate the formation and breaking of chemical bonds. It describes chemical bonds through a bond order that changes continuously with the interatomic spacing, and therefore can accurately simulate processes involving chemical reactions.

[0063] Molecular dynamics is a computer simulation method that calculates the intermolecular forces to simulate the trajectories of atoms and molecules over time. It records the trajectory q(t) of the amount of gas molecules adsorbed at a specific surface terminal. By comparing this microscopic dynamic trajectory with the macroscopic intrinsic dynamic equations (Langmuir or Linear Driving Force (LDF) model) for parameter identification, the near-wall desorption time can be calculated. Near-wall desorption time It directly reflects the intrinsic adsorption rate of the combination of the molecular structure and surface terminals of the gas to be tested at a given temperature.

[0064] In a specific example, during simulation, a periodic plate model was built using the Vienna First Principles simulation software package. The computational tool used was VASP5.4.4, and the Projector Augmented-Wave (PAW) method was used to describe the electron-ion interaction. The exchange correlation function was a Perdew-Burke-Ernzerhof (PBE) functional under the generalized gradient approximation. The plane wave cutoff energy in the basis set was 450 eV, and the Monkhorst-Pack k-point grid was 3×3×1 or adjusted according to the surface supercell size. The surface model was a combination of Si(100) crystal planes or crystal planes exposed by the actual etched cavity morphology, with 5 atomic layers and a vacuum layer >15 Å. The van der Waals correction was replaced by DFT-D3 dispersion correction. Site testing included Top, Bridge, and Hollow sites, and the adsorption energy was calculated using the site with the lowest energy. Hydroxyl, oxygen, and alkyl terminals were constructed on the surfaces of the candidate inner wall material alumina and the functionalized coating material titanium dioxide, respectively, and the adsorption energy was calculated using the site with the lowest energy.

[0065] For example, molecular dynamics theory is used to simulate and determine the desorption activation energy. When calculating the desorption parameters, the simulation temperature range is -20℃ to 150℃, and the simulation duration is no less than 2ns. Maintaining the simulation temperature within the range of -20℃ to 150℃ is used to obtain the diffusion coefficient and desorption activation energy. Temperature dependence.

[0066] By simulating the mean square displacement (MSD) at different temperatures, the relationship between the diffusion coefficient and temperature can be obtained, providing temperature-dependent parameters for signal compensation under different operating conditions (low-temperature start-up, high-temperature cleaning). The desorption activation energy is then determined. The temperature dependence can confirm the desorption activation energy. Whether the temperature changes abruptly within the target temperature range (such as the cleaning temperature) ensures the reliability of the calculation.

[0067] Optionally, the diffusion coefficient D is greater than or equal to Or, near-wall desorption time Less than or equal to 50ms.

[0068] In cleaning processes, although the vacuum pump continuously evacuates gas, a nearly stationary laminar sublayer, the near-wall concentration boundary layer, exists close to the inner wall of the detection chamber due to gas viscosity. Gas molecules adsorbed on the wall must first overcome surface adsorption forces and then pass through this stationary boundary layer to enter the mainstream gas being pumped away by the vacuum pump. If they are not completely desorbed from the near-wall concentration boundary layer, the molecules will remain "circling" near the wall, and increasing the pumping speed will not remove them; they are fixed to the wall by the boundary layer. Once the desorption distance is completed, the molecules enter the turbulent mainstream gas flow and are then pumped away by the vacuum pump.

[0069] Specifically, in gas adsorption kinetics, the near-wall desorption time The relationship between the diffusion coefficient D and the diffusion coefficient D can be shown in formula (2):

[0070] In the formula, This represents the near-wall diffusion drift distance of gas molecules as they break free from the adsorption and binding of the wall surface. 1×10 - 9 m 2 / s, then near-wall desorption time It takes 50ms.

[0071] Taking a detection chamber with a length of 20cm and a diameter of 4cm as an example, the pumping speed S of the vacuum pump is set to 2.513L / s. V represents the effective volume of the detection chamber, which is calculated to be approximately 0.2513L. The concentration of gaseous substances in the detection chamber decreases exponentially with time, and its average single-tube evacuation time... The macroscopic evacuation step and the aforementioned near-wall desorption diffusion step form a series process: based on the statistical average of the near-wall diffusion drift of a single gas molecule, the molecule first detaches from the wall after 50ms, and then is evacuated from the main detection chamber by the vacuum pump after an average single-tube evacuation time of 100ms. Together, they eliminate the gas memory effect and complete the detachment from the detection chamber in the complete cleaning response process.

[0072] When gas flows along the inner wall of the detection chamber, the velocity at the wall surface is zero, and a thin layer with a significant velocity gradient exists, called the velocity boundary layer. At the same time, gas molecules adsorbed on the wall surface must diffuse through this boundary layer before they can enter the mainstream gas flow and be pumped away by the vacuum pump. Therefore, this layer is the near-wall concentration boundary layer.

[0073] The near-wall concentration boundary layer thickness of the detection chamber ranges from 6.25 μm to 10 μm. The near-wall concentration boundary layer is a thin fluid layer that adheres to the inner wall surface of the detection chamber and has a concentration gradient. Gas molecules need to diffuse through this layer before they can enter the mainstream gas flow and be extracted by the vacuum pump. The near-wall concentration boundary layer thickness is used to characterize the equivalent distance of the near-wall mass transfer resistance concentration region.

[0074] The near-wall concentration boundary layer thickness δ is determined by the following formula (3):

[0075] In the formula, l represents the equivalent characteristic size of the detection cavity, and Sh is the Sherwood number; the equivalent characteristic size l of the detection cavity is 25 mm to 40 mm, and the Sherwood number Sh is 4000, which is determined based on the turbulent mass transfer near-wall region model.

[0076] In this invention, the detection cavity diameter is the dominant geometric dimension determining the flow state; therefore, the equivalent characteristic dimension is the cross-sectional characteristic dimension of the detection cavity, which determines the development of the mass transfer boundary layer within the detection cavity. Sh is a dimensionless number. If l = 0.04 m, Sh = 4000, and the typical lower limit of the Sherwood number for laminar mass transfer is given, the calculated near-wall concentration boundary layer thickness δ = 10 μm. This near-wall concentration boundary layer thickness of 10 μm is calculated based on the laminar mass transfer model and characterizes the diffusion distance that molecules need to traverse after desorption from the wall and into the mainstream gas. Furthermore, the minimum detection cavity diameter is 0.025 m, at which point δ = 6.25 μm.

[0077] Step S103: Select a target inner wall material from multiple candidate inner wall materials, and select a target surface terminal from multiple surface terminals.

[0078] Among them, the target adsorption energy between the target inner wall material and the molecular structure of the gas to be tested under the terminal action of the target surface satisfies the first condition, and the theoretical desorption time determined by the target desorption activation energy satisfies the second condition. The first condition is that the target adsorption energy is within -0.5 eV to -0.1 eV, and the second condition is the theoretical desorption time. Within 1 to 5 seconds.

[0079] In other words, the target inner wall material is one of the multiple candidate inner wall materials that satisfies the first condition and the second condition, the target surface terminal is one of the multiple surface terminals that satisfies the first condition and the second condition, the target adsorption energy is the adsorption energy corresponding to the target inner wall material and the target surface terminal, and the target desorption activation energy is the desorption activation energy corresponding to the target inner wall material and the target surface terminal.

[0080] The present invention selects the target inner wall material and the target surface terminal to fundamentally change the interaction type between the molecular structure of the gas to be measured and the cavity wall, transforming it from being dominated by chemical adsorption to reversible physical adsorption. The adsorption energy is controlled within a moderate range (…). The system balances adsorption and desorption, enabling the interaction between the inner wall surface of the detection chamber and the molecular structure of the gas to be tested to be dominated by reversible physical adsorption. By controlling the theoretical desorption time within 1 to 5 seconds, it can ensure that the gas molecules can be rapidly desorbed under mild heating conditions (such as 100℃~150℃) after detection.

[0081] The intermediate range ensures that gas molecules have an appropriate residence time in the optical path to guarantee detection sensitivity, while avoiding the irreversible memory effect caused by excessive adsorption and preventing insufficient gas molecule sensitivity response due to insufficient adsorption. The intermediate range avoids both strong chemisorption (… The memory effect caused by this avoids the problem of excessively weak physical adsorption. This leads to signal instability and uncontrollable desorption.

[0082] Desorption time is too fast ( Gas molecules hardly linger on the detection chamber wall, resulting in insufficient residence time and weak infrared absorption signal intensity, thus affecting the detection sensitivity of the cleaning endpoint detection device; the desorption time is too slow ( This will prolong the cleaning cycle and affect the measurement throughput. By controlling the theoretical desorption time within 1 to 5 seconds, adsorption residues can be quickly removed at the set cleaning temperature, while ensuring that gas molecules have sufficient residence time during detection, thus balancing detection sensitivity and cleaning efficiency.

[0083] Furthermore, for trace gas detection scenarios at the part-in-a-billion level, commonly found in advanced semiconductor processes, the medium range (i.e., the first condition) can be narrowed down to... To meet the signal stability requirements at lower concentrations.

[0084] After determining the target desorption activation energy, if the desorption probability of the adsorbed molecules follows an exponential decay law, the theoretical desorption time can be determined by the following formula (4):

[0085] In the formula, Represents the desorption rate constant. This indicates the frequency of attempts by surface-adsorbed molecules. , Indicates the target desorption activation energy. Represents the molar gas constant. , This indicates the target desorption temperature of the detection chamber during the cleaning process.

[0086] It characterizes the probability that adsorbed molecules successfully overcome the energy barrier and desorb from the surface per unit time. Also known as the pre-exponential factor, it characterizes the frequency at which adsorbed molecules attempt to cross the energy barrier during surface vibrations.

[0087] Step S104: Based on the target desorption activation energy, determine the target desorption temperature of the detection chamber during the cleaning process.

[0088] Specifically, after determining the two design variables—the inner wall material and the surface terminal—this invention also needs to design the target desorption temperature for the cleaning process stage of the detection chamber based on the determined target desorption activation energy. This ensures that the gas molecules adsorbed on the surface can be completely desorbed within 1 to 5 seconds at the cleaning temperature, while the maximum heating temperature is based on a structural setting that does not damage the cleaning endpoint detection device. The target desorption temperature is the thermal control threshold for the cleaning process stage, to complete the surface molecule desorption within the target desorption time.

[0089] For example, the target desorption temperature can be determined by the following formula (5):

[0090] In the formula, Indicates the target desorption temperature. This indicates the target desorption time, which theoretically ranges from 1 second to 5 seconds.

[0091] The detection chamber operates under steady-state flow conditions of continuous sample introduction and continuous gas extraction. Gas molecules enter the detection chamber from the upstream etching process and are continuously extracted by the vacuum pump, forming a stable concentration distribution and flow field within the chamber. The total residence time of a single gas molecule from entering the detection chamber to being extracted by the vacuum pump is determined by the following two sequential physical processes.

[0092] The average physical residence time of gas molecules in the main gas flow of the detection chamber depends on the ratio of the detection chamber volume V to the effective pumping speed S. In this embodiment, with the chamber geometry (length 20cm, diameter 4cm, volume approximately 0.2513L) and vacuum pump speed (2.513L / s), the average extraction time of a single tube is 100ms.

[0093] As gas molecules flow through the detection chamber, some are temporarily retained on the inner wall surface due to adsorption. After thermal desorption, they re-enter the gas phase, resulting in an additional delay relative to pure gas phase transport. This delay effect is characterized by a dimensionless delay coefficient k′, defined as the ratio of the average number of adsorbed molecules to the average number of gas phase molecules. According to adsorption thermodynamics, the delay coefficient is related to the effective adsorption energy. Satisfy the following formula (6):

[0094] In the formula, Represents Boltzmann's constant. , Indicates the total residence time of gas molecules. In one example, , .

[0095] The theoretical baseline value corresponding to an effective adsorption energy of 0.1 eV is 1.92 s. In actual processes, factors such as temperature fluctuations (±10%), vacuum pump speed reduction (≤30%), coating aging, and flow rate fluctuations can all cause response time deviations. Therefore, an engineering safety factor needs to be introduced. , The theoretical upper limit of the total residence time of gas molecules is approximately 4.8 s (1.92 × 2.5 ≈ 5 s), and the theoretical lower limit is approximately 0.28 s × 3.0 ≈ 0.84 s (1 s).

[0096] The lower limit of this window is limited by insufficient suppression of the memory effect when the adsorption energy is below 0.1 eV, while the upper limit is limited by the response delay exceeding the allowable range of the process after engineering correction of the theoretical value of 0.5 eV. Unlike the random desorption behavior of a single molecule, under steady-state conditions of continuous injection and continuous pumping throughout the detection chamber, the macroscopic calculation of the target desorption time for a large number of gas molecules ranges from 1 s to 5 s.

[0097] Step S105: Store the target desorption parameters corresponding to the target inner wall material and the target surface terminal to obtain the design parameters corresponding to the detection cavity.

[0098] Among them, the target desorption parameter is used to determine the amount of signal deviation caused by the retention of gas molecules on the inner wall during the detection stage of the detection chamber. The design parameters include the target inner wall material, target surface terminal, target desorption temperature and target desorption parameter mentioned above.

[0099] When using a clean endpoint detection device incorporating the detection chamber designed in this invention to detect gas concentration, the signal deviation (also known as signal attenuation) caused by wall adsorption retention is determined based on the target desorption parameters. Subtracting the signal deviation from the calculated concentration can eliminate residual adsorption interference to the maximum extent, achieve active suppression of baseline drift during long-term operation, and meet the needs of continuous monitoring in semiconductor production lines.

[0100] This invention can be directly based on near-wall desorption time. The signal deviation can be determined, therefore the desorption parameter can be limited to the near-wall desorption time. Alternatively, it can be based on the diffusion coefficient D and the near-wall desorption time. The signal deviation must be determined; therefore, desorption parameters may include the diffusion coefficient D and the near-wall desorption time. The specific process for determining the signal deviation based on the target desorption parameters is explained in detail below.

[0101] After determining the two design variables, inner wall material and surface terminal, this invention stores the target desorption parameters corresponding to the target inner wall material and target surface terminal, which facilitates the subsequent determination of signal deviation, overcomes the baseline drift and linearity degradation caused by adsorption, and improves the sensitivity and accuracy of concentration detection in the cleaning endpoint detection device.

[0102] The design method of the detection chamber of the cleaning endpoint detection device provided in this embodiment, after obtaining the molecular structure and operating temperature of the gas to be tested, as well as the surface energy of each candidate inner wall material under multiple different surface terminals among multiple candidate inner wall materials, selects target inner wall materials and target surface terminal combinations that meet the first and second conditions through multi-scale simulation, and modifies the inner wall of the detection chamber so that the interaction between gas molecules and the inner wall surface is dominated by reversible physical adsorption, thereby suppressing the memory effect and shortening the response time at the source.

[0103] After determining the target inner wall material and the target surface terminal, the target desorption temperature is determined based on the target desorption activation energy corresponding to the target inner wall material and the target surface terminal. This ensures that the gas molecules adsorbed on the surface can be completely desorbed within 1 to 5 seconds at the cleaning temperature. The baseline recovery speed is 5 to 10 times faster than that of traditional cavities, and the single measurement cycle can be compressed from minutes to seconds. It also avoids thermal stress damage to complementary metal-oxide-semiconductor circuits, microelectromechanical system thin film structures, and bonding interfaces on wafer-level integrated chips caused by empirical high-temperature heating.

[0104] By storing target desorption parameters corresponding to the target inner wall material and the target surface terminal for dynamic signal compensation, residual adsorption interference can be eliminated to the maximum extent during detection, achieving active suppression of baseline drift during long-term operation and controlling long-term drift within 0.5%, thus meeting the stability requirements of semiconductor process equipment for continuous operation for several months.

[0105] For example, the detection cavity can be as follows Figure 2 As shown, the detection cavity 10 includes a columnar body portion 11. The body portion 11 has a receiving cavity 111, a first opening 112 and a second opening 113. The first opening 112 and the second opening 113 are arranged opposite to each other and communicate with the receiving cavity 111. One of the first opening 112 and the second opening 113 is an input port for optical signals, and the other opening is an output port for optical signals. Figure 2 Taking the first opening 112 as the input port of the optical signal and the second opening 113 as the output port of the optical signal as an example.

[0106] The side wall of the main body 11 is also provided with a first air inlet 121 and a first air outlet 131 communicating with the receiving cavity 111. The first air inlet 121 is the input port of the gas to be measured, and the first air outlet 131 is the output port of the gas to be measured. The first air inlet 121 is located near the input port of the optical signal, and the first air outlet 131 is located near the output port of the optical signal.

[0107] Furthermore, such as Figure 2 As shown, the detection chamber also includes two extensions (a first extension 12 and a second extension 13) to facilitate connection with a transmission device for the gas to be tested. A first air inlet 121 is disposed in one of the extensions and communicates with the receiving cavity 111, and a first air outlet 131 is disposed in the other extension and communicates with the receiving cavity 111.

[0108] The present invention also provides a method for preparing a detection cavity. It should be noted that with the maturity of semiconductor processing technology, the cleaning endpoint detection device has achieved mass production compatible with complementary metal-oxide-semiconductor processes.

[0109] Clean endpoint detection devices compatible with MEMS (Micro-Electro-Mechanical Systems) are primarily manufactured using 8-inch wafer-level technology. For example, pyroelectric detectors based on scandium aluminum nitride (ScAlN) can be deposited at approximately 200°C, fully compatible with MEMS back-end processes. Furthermore, coupled strip-array waveguide structures can be realized on silicon dioxide / silicon nitride platforms using semiconductor mass production processes, providing a mass-producible technical path for on-chip integrated clean endpoint detection devices. Detection cavities can be formed on silicon wafers using semiconductor processes such as photolithography and deep reactive ion etching (DRIE), and the infrared light source, detector driving, and signal processing circuitry can be integrated onto the same substrate via a compatible MEMS platform.

[0110] Figure 3 This is a schematic flowchart of a method for preparing a detection cavity according to an embodiment of the present invention, as shown below. Figure 3 As shown, the process includes the following steps: Step S301: Provide a semiconductor structural component and form an inner wall layer with a thickness of 10 μm to 100 μm and made of the target inner wall material on the inner surface of the semiconductor structural component by a thin film growth process.

[0111] Specifically, the semiconductor structure is an undeposited detection cavity, typically made of an alloy, and the semiconductor structure is similar to... Figure 2 It includes a columnar body portion, which has a receiving cavity, a first opening and a second opening. The first opening and the second opening are arranged opposite to each other and communicate with the receiving cavity. The side wall of the body portion is also provided with a first air inlet and a first air outlet communicating with the receiving cavity.

[0112] Step S302: Place the semiconductor structure with the inner wall layer grown on the carrier plate of the process chamber.

[0113] In step S303, an atomic layer deposition process is used to alternately introduce precursors and reactants into the process chamber according to the target pulse time and target purge time, and deposit a functionalized coating on the inner wall layer to obtain the initial surface terminal.

[0114] Step S304: Water vapor is introduced into the process chamber to heat-insulate the semiconductor structure with the initial surface terminal to perform in-situ surface conditioning, so that the conditioned initial surface terminal meets the first condition and the target surface terminal is obtained.

[0115] Specifically, the process chamber is the process chamber of a semiconductor device that performs atomic layer deposition. Based on the determination of the target inner wall material and the target surface terminal during the design phase, this invention uses atomic layer deposition to deposit a functionalized coating on the inner wall of the semiconductor structure and uses water vapor to control the surface terminal in situ, thereby accurately constructing the required target surface terminal.

[0116] A gaseous source, through precursor vaporization and self-limiting chemisorption on the inner wall surface, provides the constituent elements of the functionalized coating. The reactants then react with the surface-adsorbed precursor to generate the functionalized coating. The precursor and reactants are selected based on the type of target surface terminal; for example, if the target surface terminal is a hydroxyl terminal, the precursor can be trimethylaluminum, and the reactant can be deionized water.

[0117] Atomic layer deposition (ALD) processes include multiple deposition cycles to form a functionalized coating of a predetermined thickness. A deposition cycle includes a pulse time for introducing a precursor, a purge time for introducing an inert gas, a pulse time for introducing a reactant, and a purge time for re-introducing an inert gas.

[0118] The target pulse time is the pulse time for introducing the precursor, and the target purge time is the purge time for purging with inert gas. Both the target pulse time and the target purge time are determined based on the length of the semiconductor structure and the diffusion coefficient of the precursor. For example, the target pulse time is determined by the following formula (7), and the target purge time is determined by the following formula (8).

[0119]

[0120] In the formula, The pulse time of the precursor can be represented by the minimum value determined by formula (7). The target pulse time is defined as L; L represents the near-wall diffusion drift distance of gas molecules detaching from the wall adsorption binding. Indicates the diffusion coefficient of the precursor at the process temperature; The purging time can be expressed as the minimum value determined by formula (8). The target purging time was determined.

[0121] It should be understood that atomic layer deposition (ALD) technology is renowned for its atomic-scale precision and excellent conformal coverage. In the field of advanced semiconductor packaging, ALD technology has achieved perfect conformal metal barrier layer and seed layer deposition within vias with aspect ratios exceeding 30:1, overcoming the voids and pinch-off failures caused by line-of-sight limitations in traditional physical vapor deposition (PVD) and chemical vapor deposition (CVD).

[0122] For high aspect ratio microcavity structures, thermal atomic layer deposition (ALD) can provide atomically precise, pinhole-free ultrathin films without plasma, enabling uniform coatings within nanoscale internal channels and cavities. ALD is suitable for depositing nanoscale functionalized coatings on the inner walls of detection cavities fabricated at the semiconductor wafer level.

[0123] When atomic layer deposition (ALD) is applied to vias with aspect ratios exceeding 30:1, a fixed pulse duration, such as 0.1 s, is typically used. This invention modifies the pulse duration / purge time with the detection cavity length and the diffusion coefficient of the precursor. Quantitative bonding allows precursors to penetrate centimeter-long tubes, ensuring that the entire inner wall of the semiconductor structure is covered with a functionalized coating.

[0124] Specifically, the atomic layer deposition (ALD) process comprises multiple deposition cycles divided into multiple groups, each group consisting of N deposition cycles, where N is an integer. For example, the total number of deposition cycles in the ALD process can range from 50 to 200, with N ranging from 5 to 20. For instance, a total of 50 deposition cycles could be divided into 10 groups, each consisting of 5 deposition cycles.

[0125] In the current N deposition cycles, the second air inlet of the process chamber is connected to the first opening, and the second air outlet of the process chamber is connected to the second opening, so as to form an axial airflow path in the containment chamber; in the next set of N deposition cycles, the second air inlet is connected to the second opening, and the second air outlet is connected to the first opening.

[0126] It should be understood that the current N depositional cycles and the next N depositional cycles are relative. If the current N depositional cycles are the first set of depositional cycles in the total depositional cycles, then the next N depositional cycles are the second set of depositional cycles in the total depositional cycles; if the current N depositional cycles are the second set of depositional cycles in the total depositional cycles, then the next N depositional cycles are the third set of depositional cycles in the total depositional cycles.

[0127] In other words, in the atomic layer deposition process, the direction of introduction of precursors and reactants is switched after every N deposition cycles. If the current introduction is from the first opening, the next group will be introduced from the second opening.

[0128] This invention switches the gas introduction direction every N deposition cycles, which can avoid the concentration gradient problem caused by unidirectional gas inlet in the macroscopic cavity, resulting in a thicker film at the inlet end and a thinner film at the outlet end. This ensures that the thickness uniformity deviation of the entire cavity is less than or equal to 5%, and deposits a functional coating with uniform thickness on the inner wall.

[0129] Specifically, in the initial few cycles of atomic layer deposition (ALD), the number of nucleation sites on the substrate surface is limited, and the terminal density gradually increases with the number of cycles (unsaturated growth region). After a certain number of cycles (typically 10 to 20 cycles), the surface reaches steady-state growth, and the increase in terminal density per cycle tends to be constant. In the 50 to 200 cycle stage, the film enters the steady-state growth region, and the surface terminal density reaches a stable and repeatable state. If the number of cycles is too small (e.g., <50 cycles), the film may still be in the nucleation stage, resulting in uneven terminal density and poor batch-to-batch repeatability of adsorption energy; if the number of cycles is too large (>200 cycles, >20 nm), the thickness increases but the terminal density remains unchanged, which increases stress and processing time.

[0130] For applications requiring lower adsorption energies, such as the detection of highly polar gases like nitrogen oxides, transition metal oxide coatings such as titanium dioxide and zinc oxide can be used. The surface energy state can be finely tuned by controlling the oxygen vacancy defect concentration. The surface energy state reflects the end chemical species and density, the crystallization state of the functionalized coating, and the surface profile and dangling bond density.

[0131] In one example, the substrate temperature for the atomic layer deposition process is 50°C to 300°C, and the background vacuum is less than or equal to 1 × 10⁻⁶. -3 Torr, with working pressures ranging from 0.1 Torr to 5 Torr, and target purging times ranging from 5 to 60 seconds.

[0132] Specifically, the substrate temperature is the temperature maintained by heating the semiconductor structure during the process. It is used to regulate precursor adsorption behavior and surface chemical reaction kinetics, and to maintain this temperature constant during the deposition of functionalized coatings and in-situ water vapor insulation treatment. The base vacuum is the ultimate vacuum pressure obtained by evacuating the process chamber before introducing any process gas. It is used to remove air and volatile impurities from the process chamber. The operating pressure is the total gas pressure maintained inside the process chamber when inert purge gas is continuously introduced during the atomic layer deposition cycle.

[0133] The substrate temperature in atomic layer deposition (ALD) affects the terminal density and crystallinity, which ultimately impacts the adsorption energy. If the deposition temperature is low (<50℃), the reaction between the precursor and the atoms or groups corresponding to the surface terminals may be incomplete, resulting in an excessively high terminal density, a strongly polar surface, and increased adsorption energy. If the deposition temperature is high (>300℃), the precursor may undergo thermal decomposition or desorption, and the number of atoms or groups corresponding to the surface terminals may decrease, leading to a lower terminal density.

[0134] This invention limits the substrate temperature to within 50°C to 300°C, allowing for precise control of the proportion and density of chemical species on the surface terminal and the crystallization state of the film, thereby locking the adsorption energy within a first condition (-0.5 eV to -0.1 eV). Here, the substrate refers to the underlying material supporting the film, i.e., the surface to be deposited. The substrate temperature refers to the temperature at which the substrate material is heated and maintained during the film deposition process. This temperature is typically controlled by a heating device (such as a heater, heating band, or infrared lamp) and may differ from the gas temperature within the chamber. In atomic layer deposition (ALD) processes, substrate temperature is one of the key process parameters determining deposition quality, primarily affecting: the adsorption / desorption behavior of precursors, the density, purity, and uniformity of the film, the deposition rate, and the crystal structure of the film.

[0135] For example, within the range of 50°C to 300°C, the substrate temperature selection for the atomic deposition process is related to the inner wall material of the detection cavity. The correspondence between the target inner wall material and the substrate temperature, functionalized coating, and surface terminal can be shown in Table 1.

[0136] Table 1. Correspondence between target inner wall material and substrate temperature, functionalized coating and surface terminal.

[0137] After the initial surface terminal is formed, the introduction of precursors and reactants is stopped, and water vapor is introduced into the process chamber to keep the semiconductor structure with the initial surface terminal in place for in-situ surface conditioning. This ensures that the conditioned initial surface terminal meets the first condition, thus obtaining the target surface terminal.

[0138] In one example, the water vapor pressure is 0.1 Torr to 1 Torr, and the heat treatment duration is 10 min to 30 min. If the surface terminal is hydroxyl terminal, then in-situ surface terminal control is to adjust the hydroxyl terminal density on the functionalized coating surface to a target value, which corresponds to the target adsorption energy between gas molecules and the wall surface being in the first condition (-0.5 eV to -0.1 eV).

[0139] This invention separates the step of forming the functionalized coating from the step of forming the target surface terminal. It actively controls the density of the initial surface terminal through independent water vapor post-treatment and quantitatively correlates it to the adsorption energy target window (-0.5eV to -0.1eV), so that the preparation process and the preceding design method form a complete functional closed loop.

[0140] For example, the material of the functionalized coating includes at least one of transition metal oxides such as titanium dioxide (TiO2), aluminum oxide and zinc oxide (ZnO), and the thickness of the functionalized coating is from 5 nm to 20 nm.

[0141] Functional coatings act as a physical barrier between gas molecules and the inner wall material of the chamber substrate. They must be continuous, pinhole-free, and conformally covered films to completely cover the surface of the inner wall material of the substrate below, thus completely blocking direct contact between the gas molecules to be tested. Functional coatings also serve as carriers for the target surface terminals. The thickness of the functional coating affects the surface volume and atomic number of the terminal, thereby affecting the density, distribution uniformity, and long-term stability of the terminal.

[0142] An inner wall layer with a thickness of 10 μm to 100 μm is formed on the inner surface of a semiconductor structure using either chemical vapor deposition (CVD) or physical vapor deposition (PVD). CVD is suitable for preparing high-purity, highly uniform thick films; for example, PVD silicon carbide coatings can reach 100 μm in thickness. PVD is also suitable for lower-temperature processes. The inner wall layer film exhibits good continuity, allowing for the formation of a uniform, crack-free, and pinhole-free dense coating across the entire deposition surface, preventing corrosive gases from penetrating the film and eroding the substrate. Below 10 μm, the barrier properties of the film are insufficient; above 100 μm, the internal stress of the coating increases, making it prone to cracking and peeling, and reducing the long-term reliability of the film due to stress accumulation. This thickness range maintains the bonding strength with the substrate while providing a good foundation for subsequent functionalized coatings.

[0143] The minimum thickness control for functionalized coatings is 5 nm to ensure continuity and prevent defects such as pinholes or dislocations. If the functionalized coating is discontinuous, the exposed areas of the substrate material will still undergo strong chemisorption with gas molecules, making it difficult to effectively suppress the memory effect. The maximum thickness control for functionalized coatings is 20 nm to avoid microcracks that could compromise integrity due to high internal stress, and also to prevent excessively long deposition times.

[0144] In step S305, nitrogen gas is introduced and annealing is performed to prepare the detection chamber.

[0145] Specifically, the design parameters corresponding to the detection chamber are determined by the detection chamber design method provided in any of the above embodiments. After forming the target surface terminal, the water vapor supply is stopped, and nitrogen gas is introduced into the process chamber for annealing. Specifically, after introducing nitrogen gas, the temperature is raised to 200°C to 300°C and annealed for 5 to 15 minutes. Finally, nitrogen gas is introduced to purge and cool the temperature to below 50°C, and the semiconductor structure with the target surface terminal is removed, thus obtaining the detection chamber of the clean endpoint detection device.

[0146] After forming the target surface terminal, the present invention introduces nitrogen gas and performs annealing treatment, which can stabilize the target surface terminal, remove physically adsorbed water molecules, and remove the part after cooling to avoid high-temperature oxidation and improve the preparation yield.

[0147] The detection cavity fabrication method provided by this invention forms a functionalized coating on the inner wall of a semiconductor structure through atomic layer deposition. Utilizing a self-limiting surface reaction mechanism, the coating thickness and surface terminal density can be precisely controlled at the nanoscale. By combining size-matched pulse and purge times, and alternating the precursor introduction direction between deposition cycles, a uniformly thick, conformally covered atomic-level functional layer is achieved on the macroscopic aspect ratio cavity inner wall. This effectively suppresses the adsorption and retention effect of gas molecules on the cavity inner wall, reduces the memory effect of physical adsorption, and lowers the baseline drift of the detection signal. This ensures long-term stability and repeatability in clean semiconductor cavity detection scenarios under low concentration conditions.

[0148] In some alternative embodiments, the fabrication process further includes polishing the semiconductor structure until the surface roughness of the inner surface is less than the target surface roughness.

[0149] The target surface roughness is the surface roughness determined through simulation that minimizes the adsorption site density corresponding to the surface morphology. For example, the adsorption site density under different surface morphologies can be simulated based on molecular dynamics or Monte Carlo simulations, with different surface morphologies corresponding to different surface roughnesses.

[0150] At the atomic scale, an ideal flat surface has a regular arrangement of atoms. However, in actual processing, the etched surface will inevitably have defects such as steps, kinks, and vacancies. The atoms at the defect sites have unsaturated coordination numbers, a large number of dangling bonds, and lower local potential energy, resulting in extremely strong interactions with gas molecules.

[0151] By simulating the density of adsorption sites (i.e., defect sites) under different surface morphologies, the surface morphology can be converted into the number of adsorption sites. By selecting the lowest adsorption site density, defects such as steps / bends are minimized, making the actual inner wall surface approach an ideal smooth surface. The adsorption behavior of gas molecules is mainly determined by the surface terminals. At this point, the actually measured adsorption energy will be consistent with the target adsorption energy, ensuring that it falls precisely within the medium range of -0.5 eV to -0.1 eV.

[0152] This embodiment controls the surface roughness of the inner wall of the detection chamber to be below the target surface roughness, which can avoid unpredictable strong adsorption caused by physical defects, ensure that the target surface terminal becomes the only dominant adsorption factor, lock the final adsorption energy in the medium range, and thus suppress the memory effect caused by strong chemisorption.

[0153] The present invention also provides a cleaning endpoint detection device, such as... Figure 4As shown, the cleaning endpoint detection device 100 includes an infrared light source 20, an infrared detector 30, a controller 40, and a detection cavity 10. The detection cavity 10 is prepared using the detection cavity preparation method provided in any of the above embodiments.

[0154] Specifically, the infrared light source 20 is used to generate a first optical signal of the target wavelength and a second optical signal of the reference wavelength. The gas to be measured is introduced into the detection cavity 10. The first and second optical signals are transmitted to the infrared detector 30 after passing through the detection cavity 10. The first optical signal is absorbed to different degrees according to the concentration change of the gas to be measured.

[0155] Infrared detector 30 is used to convert the received first optical signal into a first electrical signal and the received second optical signal into a reference electrical signal. Controller 40 is used to determine the absorptivity based on the first electrical signal and the reference electrical signal to obtain the concentration of the gas to be measured. The method by which the controller determines the concentration using the above formula (1), the first electrical signal, and the reference electrical signal can be found in relevant technologies and will not be described in detail here.

[0156] Furthermore, such as Figure 5 As shown, the cleaning endpoint detection device 100 also includes a temperature sensing module 50, and the controller 40 stores target desorption parameters, which are determined by the design method of the detection cavity of the cleaning endpoint detection device provided in any of the above embodiments.

[0157] Specifically, the controller 40 is also used to obtain the current temperature of the detection chamber 10 from the temperature sensing module 50, and to obtain the measured concentration of the gas to be tested in the detection chamber; the controller 40 is also used to construct a kinetic model based on the current temperature, the measured concentration and the target desorption parameters, and to determine the signal deviation caused by the retention of gas molecules on the inner wall of the detection chamber through the kinetic model; the controller 40 is also used to correct the deviation of the measured concentration of the gas to be tested based on the signal deviation, and to obtain the corrected concentration.

[0158] For continuous flow systems like the detection chamber, the kinetic model uses the axial diffusion model (ADM) coupled with the wall adsorption term, as shown in equation (9).

[0159]

[0160] In the formula, Indicates the time of the gas being tested. and spatial coordinates Instantaneous concentration at the location, Indicates the effective diffusion coefficient. The diffusion coefficient D is the value corrected for the current temperature. This represents the spatial coordinates along the axial direction of the detection cavity. The flow rate of the gas being measured is obtained from the mass flow controller. This represents the total surface area of ​​the cavity's inner wall. Indicates the effective volume of the detection cavity. This represents the adsorption rate constant. , This represents the equilibrium concentration at the current temperature and gas partial pressure, for a specific combination of the analyte gas and the surface terminal. Determined by the target adsorption energy, the more negative the target adsorption energy, the stronger the adsorption. The lower.

[0161] Reading the flow rate of the gas to be measured Current temperature T and measured concentration Then, the pre-stored diffusion coefficient D and near-wall desorption time are called. And derive based on the target adsorption energy Then, along the cavity axis z from 0 to L, the above formula (9) is solved using numerical methods such as the finite difference method to obtain the concentration distribution at any time and any position in the entire cavity. The cleaning endpoint detection device measures the integral average concentration along the entire optical path. The concentration at an integral average, rather than the concentration at a single point. It can be determined by the following formula (10):

[0162] Determining the integral average concentration Then, the difference between the measured concentration and the integrated average concentration can be used as the signal deviation, i.e., the signal deviation is shown in formula (11):

[0163] In the formula, This indicates the amount of signal deviation.

[0164] After determining the signal deviation, the sum of the signal deviation and the concentration of the gas to be measured can be used as the corrected concentration of the gas to be measured, as shown in formula (12):

[0165] In the formula, This represents the corrected concentration, i.e., the concentration value after wall adsorption compensation. The concentration calculated based on the first electrical signal and the reference electrical signal is C in the above formula (1).

[0166] This embodiment uses numerical solution and integration along the optical path to accurately calculate the deviation between the apparent concentration measured by the cleaning endpoint detection device and the measured concentration. This enables the real-time subtraction of concentration decay caused by wall adsorption from the original signal, thus achieving active suppression of baseline drift.

[0167] It should be noted that the diffusion coefficient D describes the migration rate of gas molecules in the bulk cavity (gas phase), and the near-wall desorption time. The rate at which gas molecules are captured by the wall surface and reach adsorption equilibrium is described by formula (13), where the diffusion coefficient D and the near-wall desorption time are also considered. Together they determine the overall rate of the adsorption process.

[0168]

[0169] In the formula, Indicates the total adsorption time. Indicates diffusion time, proportional to ,Right now .

[0170] When the cavity size is very small ( ), much smaller The diffusion process is much faster than the adsorption process; molecules reach the wall almost instantaneously, and the total desorption rate can be increased by... When controlled individually, the axial concentration distribution tends to be uniform. ), diffusion term The above formula (9) can be simplified to the simplified form shown in formula (14).

[0171]

[0172] At this point, the diffusion coefficient D can be solved without further calculation; the desorption parameter can be simply the near-wall desorption time. .

[0173] For extremely small microcavities, the diffusion path is very short, diffusion is faster than adsorption, axial concentration uniformity is good, and the contribution of the axial diffusion term is negligible.

[0174] For example, if Using the kinetic model shown in formula (14), the desorption parameter is only the near-wall desorption time. Since the diffusion coefficient D is not required, the computational complexity is reduced; if The kinetic model shown in formula (9) is adopted, and the desorption parameters include the diffusion coefficient D and the near-wall desorption time. This improves computational accuracy and ensures versatility.

[0175] Optionally, such as Figure 5 As shown, the cleaning endpoint detection device 100 also includes a heating module 60.

[0176] The controller 40 is also used to control the heating module 60 to heat the detection chamber before detection, so as to desorb the gas molecules adsorbed on the inner wall of the detection chamber to eliminate the memory effect; the target desorption temperature is determined by the design method of the detection chamber of the cleaning endpoint detection device provided in any of the above embodiments; the controller 40 is also used to acquire the detection signal of the detection chamber during the heating desorption process, and when the detection signal recovers from the stable concentration value along the falling edge to the clean state baseline value, extract the time interval of the falling edge of the detection signal as the actual desorption time; and adjust the current heating temperature of the heating module 60 according to the deviation between the acquired actual desorption time and the target desorption time, and perform heating according to the adjusted current heating temperature.

[0177] The target desorption time is the total equivalent time from the start of heating in the detection chamber to the point where wall adsorption, molecular desorption, and removal by the vacuum pump, and the detection signal recovers to the baseline value of the clean state, including the near-wall molecular desorption time and the gas replacement time during continuous gas intake. The gas replacement time is represented by the average extraction time per tube. The near-wall molecular desorption time is the characteristic time required for gas molecules to move from being adsorbed on the inner wall of the detection chamber, desorb, and cross the near-wall concentration boundary layer before entering the bulk gas phase. The average extraction time per tube is the average time required for gas molecules to be extracted from the detection chamber by the vacuum pump after entering the bulk gas phase; its value is determined by the ratio of the detection chamber volume to the effective pumping speed of the vacuum pump. Under continuous flow conditions with continuous gas intake, the macroscopic extraction process consists of multiple microscopic extraction events of gas from multiple single tubes. The target desorption time is the equivalent amplified value of the sum of the near-wall molecular desorption time and the average extraction time per tube.

[0178] The method for real-time adjustment of the current temperature is based on the advance adjustment of the signal change rate. During the heating and desorption process, the controller samples the detection signal at fixed time intervals (e.g., 10ms) and calculates the signal change rate as shown in formula (15):

[0179] In the formula, r(t) represents the rate of signal recovery at the current moment, S(t) is the value of the detected signal at the current moment, and Δt is the sampling interval, i.e., the fixed time interval.

[0180] When the signal recovery rate slows down, the temperature is raised immediately without waiting for the signal to lag significantly. The slowdown means that desorption is about to slow down. Adjusting the temperature in advance for real-time dynamic temperature control can compensate for the delay caused by thermal inertia before the actual slowdown.

[0181] For example, during the heating desorption process, the controller acquires the detection signal at a fixed sampling interval of 10ms and calculates the signal change rate r(t) in real time to determine the trend of the desorption rate. Taking an initial heating temperature of 130℃ as an example, when desorption proceeds normally, the baseline signal value in the clean state is 5.0V, and the signal value corresponding to gas adsorption after an interval of 0.8s is 4.2V. During this process, when the total signal change range is 0.8V, the calculated signal change rate r(t) stabilizes at around 1.0V / s.

[0182] When the controller detects a continuous decrease in r(t) (e.g., from 1.02V / s to 0.95V / s), it determines that the desorption rate is slowing down and immediately issues a temperature increase command, raising the target temperature by 2K~3K. After the temperature is increased, if the downward trend of r(t) stops, the current temperature is maintained; if it continues to decrease, it is increased again by 2K~3K; if r(t) remains too high, the temperature is appropriately lowered. Throughout the process, the controller continuously decides whether to increase, decrease, or maintain the temperature based on the real-time trend of r(t), with the temperature fluctuating around the initial set value, typically within ±10K. Through this successive fine-tuning method, the temperature is always kept near the dynamic equilibrium point that meets the target desorption time requirement, causing the actual desorption time to converge to the target desorption time range.

[0183] The prepared detection chamber is installed on the cleaning endpoint detection device. Baseline measurement and zero-point calibration are performed before detecting the gas concentration. In the preheating cleaning process stage, before the cleaning process detection, the residual gas molecules adsorbed on the inner wall are fully desorbed. The initial cleaning temperature is the target desorption temperature, which is determined by formula (5), or the product of the target desorption temperature and the weighting coefficient. After the temperature of the detection chamber stabilizes, nitrogen gas is introduced. The absorption signal obtained by the infrared detector 30 under the nitrogen atmosphere based on the received infrared light intensity conversion is used as the initial value of the reference electrical signal. After stabilization for a period of time, zero-point calibration is performed.

[0184] During the cleaning process testing, a suitable current heating temperature for the gas to be tested is set, and heating is performed at this temperature. An experimental environment is established so that the detection chamber is continuously filled with a concentration of 500 ppm, stabilized for 20 seconds, and then the concentration is set to 0. During the heating desorption process, the controller continuously acquires detection signals at fixed sampling intervals, records the concentration curve in real time, and extracts the time interval corresponding to the recovery of the original stable concentration detection signal to the clean state baseline value, which is the actual desorption time for this heating desorption.

[0185] The actual desorption time characterizes the time it takes for the concentration of the wall adsorbent layer to decrease to the target level, rather than the entire heating process duration, and its value is usually less than the total duration of the cleaning process. Microscopically, this actual desorption time includes the sum of the near-wall molecule desorption time and the average extraction time per tube. After being amplified by the adsorption retention factor, it becomes a macroscopically equivalent time comparable to the target desorption time (1s~5s). The comparison between the two is used to detect the real-time adjustment of the chamber temperature.

[0186] At the microscopic level, during the heating desorption process, the microscopic physical composition consists of the sum of the near-wall molecule desorption time (approximately 50 ms) and the average single-tube extraction time (approximately 100 ms). Ideally, the theoretically required time for a tube of gas to desorb from the wall and completely exit the detection chamber is approximately 150 ms. At the macroscopic level, in a continuously flowing system with continuous gas intake, due to the delayed effect of the wall adsorption-desorption dynamic equilibrium, this microscopic time needs to be multiplied by the adsorption-retention factor (3~10) and then amplified by an engineering safety factor (2.5~3.0) to obtain a macroscopically equivalent desorption time comparable to the target desorption time (1s~5s). The equivalent time constant extracted by the controller from the signal recovery curve is this macroscopically equivalent time, and its physical composition includes the combined effects of adsorption retention and engineering amplification.

[0187] The heating module dynamically heats the device based on real-time adjustments to the current heating temperature to ensure efficient desorption. The entire temperature regulation process is completed within the same temperature control step. Once the current cleaning process is finished, the detection chamber is cooled to the operating temperature, nitrogen is introduced, and the reference electrical signal is confirmed to have returned to its initial value. In subsequent cleaning processes, the controller uses a real-time adjustment strategy to accumulate historical experience and repeats the above detection process.

[0188] After prolonged contact with trace amounts of organic, acidic, and alkaline contaminants, the active sites on the inner wall surface of the detection chamber will slowly change, causing continuous drift in molecular adsorption / desorption kinetics. This embodiment uses baseline recovery time as a quantitative indicator to sense the coating's contamination and aging levels in real time. By adjusting the current heating temperature according to the baseline recovery time, it can autonomously correct baseline drift, reducing the impact of drift after long-term operation. This ensures that even when the desorption kinetics of the coating surface change due to contamination or oxidation during long-term sensor operation, optimal cleaning efficiency and low power consumption can still be maintained through parameter self-adjustment.

[0189] For example, controller 40 is used to dynamically increase the current heating temperature when the actual desorption time is greater than the target desorption time; controller 40 is used to dynamically decrease the current heating temperature when the actual desorption time is less than the target desorption time; controller 40 is used to control the current heating temperature to remain unchanged when the actual desorption time is equal to the target desorption time.

[0190] Specifically, after the controller obtains the actual desorption time of gas molecules in the detection chamber, it determines the deviation ratio using formula (16), and then adjusts the current heating temperature of the heating module according to the deviation ratio.

[0191]

[0192] In the formula, p represents the deviation ratio, and t actual t represents the actual desorption time of gas molecules within the detection chamber. target Indicates the time of target desorption.

[0193] The specific process by which the controller adjusts the current heating temperature of the heating module according to the deviation ratio can be as follows: when t actual >t target When p > 0, the controller will increase the current heating temperature by K × p × T. step The adjusted current heating temperature is: T adjust =T current +K×p×T step When t actual <t target When p < 0, the controller will reduce the current heating temperature by K × |p| × T. step The adjusted current heating temperature is: T adjust =T current -K×|p|×T step When t actual =t target When p=0, the controller maintains the current heating temperature.

[0194] T current T represents the current heating temperature of the heating module. adjust T represents the adjusted current heating temperature. step This indicates the preset temperature step size, and K is the gain coefficient. The gain coefficient is a preset constant or determined by the controller based on the deviation ratio. The value of K ranges from 1 to 8.

[0195] Specifically, the deviation ratio p represents the relative degree to which the current desorption time deviates from the target value. The larger the value, the more severe the deviation, and the greater the temperature adjustment required. The adjustment range is determined by the deviation ratio, and the adjustment direction is determined by the deviation direction. Through continuous feedback, the temperature automatically converges to the optimal value that allows desorption to be completed within the target time.

[0196] For example, when a single over-adjustment causes the actual desorption time to exceed the target value (from slow to fast, or from fast to slow), the controller detects a reversal of the deviation direction in the next round of sampling and automatically performs reverse adjustment to bring the system into dynamic equilibrium near the target value.

[0197] The gain coefficient K can be adaptively adjusted by the controller according to the deviation convergence trend. For example, K is increased to enhance the adjustment force when convergence is too slow, K is decreased to suppress overshoot when oscillation intensifies, and K is kept unchanged when convergence is normal. The smaller the value, the closer it is to the target, and the smaller the adjustment amount is required; when the value is zero, no adjustment is required.

[0198] Controller 40 obtains the actual desorption time If t actual >t target If the actual desorption time is too slow, it indicates that the current temperature is too low, resulting in insufficient desorption rate. In this case, the controller will increase the temperature. actual <t target If the actual desorption time is too fast, it indicates that the current temperature is too high, and the controller will lower the temperature.

[0199] The adjusted current heating temperature range is ±10℃ to avoid affecting the basic desorption efficiency. K is the gain coefficient, which determines the adjustment level; T step The preset temperature step size determines the base amplitude of a single adjustment; the product of the two and the deviation ratio together determine the actual adjustment amount, so that the adjustment amplitude always adapts to the degree of deviation.

[0200] It should be understood that theoretical desorption time characterizes the theoretical value of the expected completion time of the desorption process calculated based on the adsorption-desorption kinetic model (including physical parameters such as adsorption energy, temperature, and diffusion coefficient), reflecting the mathematical prediction result of the physical model. Target desorption time, on the other hand, characterizes the preset benchmark value used by the controller to compare with the detected actual desorption time in actual process control, reflecting the process control objective that the control system aims to achieve. The former emphasizes the source of theoretical calculations, while the latter emphasizes actual control.

[0201] The present invention also provides a gas concentration measurement method for a cleaning endpoint detection device, wherein the cleaning endpoint detection device is any of the cleaning endpoint detection devices provided in the above embodiments, and the gas concentration measurement method can be used in the controller of the cleaning endpoint detection device. Figure 6 This is a schematic flowchart of a gas concentration measurement method using a cleaning endpoint detection device according to an embodiment of the present invention, as shown below. Figure 6 As shown, the process includes the following steps: Step S601: The heating module is heated according to the current heating temperature to desorb the gas molecules remaining on the inner wall of the detection chamber.

[0202] Specifically, the current heating temperature is a preset temperature suitable for detecting the gas to be detected.

[0203] In step S602, after the current temperature of the detection chamber drops to the operating temperature, nitrogen gas is introduced into the detection chamber, the infrared detector detects the reference electrical signal, and zero-point calibration is performed.

[0204] Specifically, the reference electrical signal obtained during the baseline measurement stage is the baseline under zero gas conditions. It generally needs to be heated stably for 30 to 60 minutes until the current baseline is stable. Then, click the reset button to perform zero-point calibration.

[0205] Step S603: The gas to be tested is introduced into the detection cavity, and the infrared light source is controlled to generate a first optical signal of the target wavelength and a second optical signal of the reference wavelength.

[0206] The first optical signal and the second optical signal are transmitted to the infrared detector after passing through the detection cavity. The first optical signal is absorbed to different degrees according to the concentration change of the gas to be measured. The infrared detector is used to convert the received first optical signal into a first electrical signal and the received second optical signal into a reference electrical signal.

[0207] Step S604: Determine the absorption rate based on the first electrical signal and the reference electrical signal to obtain the concentration of the gas to be measured.

[0208] Specifically, the method for determining the concentration based on the first electrical signal and the reference signal is based on relevant technologies and will not be described in detail here.

[0209] Step S605: Obtain the current temperature of the detection chamber and the measured concentration of the gas to be tested in the detection chamber.

[0210] Specifically, the current temperature is obtained from the temperature sensing module and the measured concentration is obtained from the mass flow controller.

[0211] Step S606: Based on the kinetic model constructed using the current temperature, measured concentration, and target desorption parameters, determine the signal deviation caused by the retention of gas molecules on the inner wall of the detection chamber.

[0212] Specifically, the signal deviation can be determined by combining the above formulas (9) to (11). .

[0213] Step S607: Correct the deviation of the measured concentration of the gas to be measured according to the signal deviation to obtain the concentration of the gas to be measured.

[0214] Specifically, the concentration of the gas to be measured can be determined using the above formula (12).

[0215] For example, the following describes the real-time temperature adjustment process of the detection chamber in the detection of a cleaning process.

[0216] Stage Zero: Preheating of the detection chamber (independent of the cleaning process); Before the cleaning process is started, the detection chamber has independently completed preheating and stabilized at a set value of 130°C, and is in a state to be detected. Thereafter, throughout the entire cleaning process, the temperature of the detection chamber is dynamically adjusted on the basis of 130°C according to the feedback of the actual desorption time.

[0217] Phase 1: Process chamber heating and preheating (0s~30s, the detection chamber is maintained at 130°C); The process chamber is heated from 80°C to 130°C, no cleaning gas is introduced yet, but pure nitrogen is introduced. The detection chamber is maintained at 130°C, the signal is at a clean state baseline value of 5.0V, zero calibration is performed, and the controller continuously monitors.

[0218] Phase 2: Main cleaning phase (30s~90s, the temperature of the detection chamber starts to be dynamically adjusted); After the cleaning gas is introduced and the plasma is turned on, the signal of the detection chamber starts to change. The controller measures the actual desorption time t for the first time actual =2.8s (relatively slow), and the temperature of the detection chamber is increased from 130°C to 132.4°C. Subsequently, as the reaction process in the process chamber advances, the temperature of the detection chamber dynamically fluctuates between 130°C and 134°C, the adjustment amplitude of each round is determined by the deviation ratio p, and the controller continuously decides to heat up, cool down or maintain the temperature in each sampling cycle.

[0219] This adjustment process is based on the target desorption time, takes the deviation ratio between the actual desorption time and the target value as the adjustment basis, and gradually pushes the temperature to the optimal point that enables desorption to be completed within the target time through proportional feedback. When the deviation tends to zero, the temperature tends to be stable; when the deviation reappears, the adjustment restarts. This control logic does not rely on specific process parameters, and is applicable to any thermal desorption process that takes desorption time feedback as the control target. The adjustment amplitude always dynamically changes following the magnitude of the deviation, and the direction is switched in real time following the sign of the deviation, ensuring that the system converges near the target value without divergence.

[0220] The controller automatically adjusts the K value according to the deviation ratio p: when p>0.3, the deviation is large, and the K value is increased to 5~8 for fast response; when 0.1<p≤0.3, the deviation is medium, and the K value is maintained in the range of 3~5; when p≤0.1, the deviation is small, and the K value is decreased to 1~3 to avoid overshoot. The controller compares the deviation changes after two consecutive adjustments: if the reduction amplitude of the deviation ratio exceeds 30%, it indicates that the current K value is appropriate and remains unchanged; if the reduction amplitude is less than 20%, it indicates that the adjustment force is insufficient, and the K value is increased by 0.5~1; if the deviation ratio does not decrease but increases, it indicates that over-adjustment causes oscillation, and the K value is decreased by 0.5~1. The change amplitude of the K value in a single time shall not exceed 0.5~1, and the change direction can only be executed after two consecutive confirmations, so as to avoid severe temperature fluctuation caused by sudden change of the K value. The K value is always limited to the safe range of 1~8.

[0221] Phase 3: Cleaning endpoint determination (90s~100s); Byproducts on the inner wall of the process chamber were largely removed, and the detection chamber signal returned to the cleanliness baseline value of 5.0V. The controller determined that the cleaning endpoint had been reached, stopped temperature regulation, and the detection chamber temperature stabilized at approximately 133°C.

[0222] Phase 4: Purge and cooling (100s~120s); The process chamber stops reacting and is purged again with N2. The temperature in the detection chamber naturally drops from 133℃ or is actively cooled to the standby temperature, awaiting the next cleaning process detection.

[0223] The gas concentration measurement method of the cleaning endpoint detection device provided in this embodiment has an inner wall surface of the detection chamber modified by atomic layer deposition, so that the interaction between gas molecules and the inner wall surface is dominated by reversible physical adsorption, the target adsorption energy is in the medium range, and the gas molecules can be rapidly desorbed under mild heating conditions after the detection is completed.

[0224] The cleaning temperature in the preheating cleaning stage is determined based on the target desorption activation energy, which avoids blindly heating at high temperatures. The cleaning temperature is controlled between 100℃ and 150℃, reducing power consumption by more than 50% and preventing thermal stress damage to wafer-level integrated chips from extreme thermal shocks. Using the simulated target desorption parameters for dynamic signal compensation, the concentration attenuation caused by wall adsorption can be subtracted from the original signal in real time, thus achieving active suppression of baseline drift.

[0225] Meanwhile, during the long-term operation of the cleaning endpoint detection device, the baseline recovery time is used as a quantitative indicator to sense the coating contamination and aging degree in real time. The current heating temperature is adjusted in real time according to the baseline recovery time, and the baseline drift is automatically corrected to keep the long-term drift within 0.5%. This ensures that when the desorption kinetics of the coating surface change due to contamination or oxidation, the device can still maintain the optimal cleaning efficiency and low power consumption characteristics through parameter self-adjustment.

[0226] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0227] like Figure 7 As shown, the electronic device includes a memory 710 and a processor 720. The memory 710 stores a computer program, and the processor 720 is configured to run the computer program to execute the design method of the detection chamber of any of the above-described cleaning endpoint detection devices, or to execute the steps in the gas concentration measurement method embodiments of any of the above-described cleaning endpoint detection devices. Figure 7 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0228] Furthermore, the electronic device also includes a communication interface 730 for communicating with other devices or communication networks. Figure 7 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0229] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the design method of the detection chamber of the cleaning endpoint detection device or the gas concentration measurement method of the cleaning endpoint detection device shown in the above embodiments is implemented.

[0230] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0231] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for designing a detection cavity, the detection cavity being used to introduce a test gas to detect the cleaning endpoint of a semiconductor process chamber, characterized in that, The design method includes: The molecular structure, operating temperature, and surface energy of each candidate inner wall material under multiple different surface terminals are obtained from the test gas. The surface terminal is the atomic layer that comes into contact with the test gas after the inner wall of the detection cavity is functionalized by an atomic layer deposition process. At the operating temperature, the adsorption energy, desorption activation energy, and desorption parameters between the molecular structure of the gas under test and the surface terminals corresponding to multiple surface energies are simulated and determined. The desorption parameters include the diffusion coefficient and the near-wall desorption time determined by the diffusion coefficient. A target inner wall material is selected from a plurality of candidate inner wall materials, and a target surface terminal is selected from a plurality of surface terminals. The adsorption energy between the target inner wall material and the molecular structure of the gas to be tested under the action of the target surface terminal meets a first condition, and the theoretical desorption time determined by the desorption activation energy meets a second condition. The adsorption energy that meets the first condition is taken as the target adsorption energy, and the desorption activation energy that meets the second condition is taken as the target desorption activation energy. The first condition is that the target adsorption energy is within -0.5 eV to -0.1 eV, and the second condition is that the theoretical desorption time is within 1 s to 5 s. Based on the target desorption activation energy, the target desorption temperature of the detection chamber during the cleaning process is determined; The desorption parameters corresponding to the target inner wall material and the target surface terminal are stored as target desorption parameters to obtain the design parameters corresponding to the detection cavity. The target desorption parameters are used to determine the signal deviation caused by gas molecules being retained on the inner wall during the detection stage. The design parameters include the target inner wall material, the target surface terminal, the target desorption temperature, and the target desorption parameters.

2. The design method according to claim 1, characterized in that, The surface terminal is a hydrogen terminal, a hydroxyl terminal, an oxygen terminal, a chlorine terminal, an alkyl terminal, or a fluoroalkyl terminal.

3. The design method according to claim 1, characterized in that, The candidate inner wall materials are silicon nitride, yttrium oxide, silicon carbide, stainless steel, aluminum oxide, or polytetrafluoroethylene.

4. The design method according to any one of claims 1 to 3, characterized in that, The near-wall concentration boundary layer thickness of the detection chamber ranges from 6.25 μm to 10 μm. This near-wall concentration boundary layer is a fluid layer adhering to the inner wall surface of the detection chamber and exhibiting a concentration gradient. The thickness of the near-wall concentration boundary layer characterizes the equivalent distance of the near-wall mass transfer resistance concentration region. The thickness of the near-wall concentration boundary layer is determined by the following formula: δ=l / Sh; In the formula, δ represents the thickness of the near-wall concentration boundary layer, l is the equivalent characteristic size of the detection cavity, Sh is the Sherwood number, and the equivalent characteristic size is the cross-sectional characteristic size of the detection cavity.

5. The design method according to any one of claims 1 to 3, characterized in that, Determining the target desorption temperature of the detection chamber during the cleaning process based on the target desorption activation energy includes: The target desorption temperature is determined using the following formula: ； In the formula, Indicates the target desorption temperature. Indicates the target desorption activation energy. Represents the gas constant. , This indicates the frequency of attempts by surface-adsorbed molecules. The target desorption time is indicated, and the target desorption temperature is the thermal control threshold of the cleaning process stage, so as to complete the surface molecule desorption within the target desorption time.

6. The design method according to any one of claims 1 to 3, characterized in that, The diffusion coefficient is greater than or equal to Alternatively, the near-wall desorption time is less than or equal to 50 ms.

7. A method for preparing a detection cavity, characterized in that, The preparation method includes: A semiconductor structural component is provided, wherein an inner wall layer with a thickness of 10 μm to 100 μm and made of a target inner wall material is formed on the inner surface of the semiconductor structural component by a thin film growth process. The semiconductor structural component includes a columnar body portion, the body portion having a receiving cavity, a first opening and a second opening, the first opening and the second opening being disposed opposite to each other and communicating with the receiving cavity, and the sidewall of the body portion also having a first air inlet and a first air outlet communicating with the receiving cavity. A semiconductor structure with an inner wall layer is placed on a carrier plate in a process chamber. In the current N deposition cycles, the second air inlet of the process chamber is connected to the first opening, and the second air outlet of the process chamber is connected to the second opening to form an axial airflow path within the receiving cavity. In the next set of N deposition cycles, the second air inlet is connected to the second opening, and the second air outlet is connected to the first opening, where N is an integer. An atomic layer deposition process is used to alternately introduce a precursor and a reactant into the process chamber according to a target pulse time and a target purge time, and a functionalized coating is deposited on the inner wall layer to obtain an initial surface terminal. The target pulse time and the target purge time are both determined based on the length of the semiconductor structure and the diffusion coefficient of the precursor. Water vapor is introduced into the process chamber to heat-insulate the semiconductor structure on which the initial surface terminal is formed, so as to perform in-situ surface conditioning and make the conditioned initial surface terminal meet the first condition to obtain the target surface terminal. The detection cavity is prepared by introducing nitrogen gas and annealing, wherein the design parameters of the detection cavity are determined by the design method of the detection cavity according to any one of claims 1 to 6.

8. The preparation method according to claim 7, characterized in that, The material of the functionalized coating includes at least one of titanium dioxide, aluminum oxide and zinc oxide, and the thickness of the functionalized coating is 5 nm to 20 nm.

9. The preparation method according to claim 7, characterized in that, The substrate temperature for the atomic layer deposition process is 50°C to 300°C, and the base vacuum is less than or equal to 1×10⁻⁶. -3 The working pressure is 0.1 Torr to 5 Torr, the target purging time is 5s to 60s, the total number of deposition cycles is 50 to 200, and the value of N is 5 to 20.

10. The preparation method according to any one of claims 7 to 9, characterized in that, The pressure of the water vapor is 0.1 Torr to 1 Torr, and the duration of the heat preservation treatment is 10 min to 30 min.

11. A cleaning endpoint detection device, characterized in that, The cleaning endpoint detection device detects the concentration change of the gas to be tested using the nondispersive infrared principle to provide feedback on the cleaning endpoint. The cleaning endpoint detection device includes an infrared light source, an infrared detector, a controller, and a detection chamber, wherein the detection chamber is prepared by the method of any one of claims 7 to 10. The infrared light source is used to generate a first light signal at the target wavelength and a second light signal at the reference wavelength. The gas to be tested is introduced into the detection cavity. The first light signal and the second light signal are transmitted to the infrared detector after passing through the detection cavity. The gas to be tested absorbs the first light signal. The infrared detector is used to convert the received first optical signal into a first electrical signal and the received second optical signal into a reference electrical signal; The controller is used to determine the absorption rate based on the first electrical signal and the reference electrical signal to obtain the concentration of the gas to be measured.

12. The cleaning endpoint detection device according to claim 11, characterized in that, The cleaning endpoint detection device also includes a temperature sensing module, and the controller stores target desorption parameters. The controller is also used to obtain the current temperature of the detection chamber from the temperature sensing module, and to obtain the measured concentration of the gas to be tested in the detection chamber; A kinetic model is constructed based on the current temperature, the measured concentration, and the target desorption parameters. The signal deviation caused by the retention of gas molecules on the inner wall of the detection chamber is determined by the kinetic model. The deviation in the measured concentration of the gas to be measured is corrected based on the signal deviation to obtain the corrected concentration.

13. The cleaning endpoint detection device according to claim 12, characterized in that, The cleaning endpoint detection device further includes a heating module, and the controller is used for: Before detection, the heating module is controlled to heat the detection chamber, causing the gas molecules adsorbed on the inner wall of the detection chamber to desorb, thereby eliminating the memory effect; During the heating desorption process, the detection signal of the detection chamber is acquired. When the detection signal recovers from the stable concentration value along the falling edge to the clean state baseline value, the time interval of the falling edge of the detection signal is extracted as the actual desorption time. The heating temperature of the heating module is adjusted based on the deviation between the actual desorption time and the target desorption time. The target desorption time is the total equivalent time from the start of heating of the detection chamber to wall adsorption, molecular desorption and removal by the vacuum pump, and the recovery of the detection signal to the baseline value of the clean state. It includes the near-wall molecular desorption time and the continuous air intake and gas replacement time. The gas replacement time is represented by the average extraction time of a single tube.

14. The cleaning endpoint detection device according to claim 13, characterized in that, The controller determines the deviation ratio using the following formula: p=(t actual -t target ) / t target ; In the formula, p represents the deviation ratio, and t actual t represents the actual desorption time of gas molecules within the detection chamber. target Indicates the target desorption time; The controller adjusts the current heating temperature of the heating module according to the deviation ratio; wherein, when t actual >t target When p>0, the controller increases the current heating temperature by K×p×T. step The adjusted current heating temperature is: T adjust =T current +K×p×T step When t actual <t target When p < 0, the controller will reduce the current heating temperature by K × |p| × T. step The adjusted current heating temperature is: T adjust =T current -K×|p|×T step When t actual =t target When p=0, the controller maintains the current heating temperature unchanged; T current T represents the current heating temperature of the heating module. adjust T represents the adjusted current heating temperature. step The preset temperature step size is represented by K, which is the gain coefficient. The gain coefficient is a preset constant or determined by the controller according to the deviation ratio, and the value of K ranges from 1 to 8.

15. A method for measuring gas concentration using a clean endpoint detection device, characterized in that, The cleaning endpoint detection device is the cleaning endpoint detection device according to any one of claims 11 to 14, and the gas concentration measurement method is applied to the controller of the cleaning endpoint detection device, the gas concentration measurement method comprising: Heating is controlled according to the current heating temperature to heat the heating module, causing the gas molecules remaining on the inner wall of the detection chamber to desorb; After the current temperature of the detection chamber drops to the operating temperature, nitrogen gas is introduced into the detection chamber, and the infrared detector detects the reference electrical signal and performs zero-point calibration. A gas to be tested is introduced into a detection cavity, and an infrared light source is controlled to generate a first light signal of a target wavelength and a second light signal of a reference wavelength. The first light signal and the second light signal are transmitted to an infrared detector after passing through the detection cavity. The gas to be tested absorbs the first light signal. The infrared detector is used to convert the received first light signal into a first electrical signal and the received second light signal into a reference electrical signal. The absorption rate is determined based on the first electrical signal and the reference electrical signal to obtain the concentration of the gas to be measured; The current temperature of the detection chamber and the measured concentration of the gas to be tested in the detection chamber are obtained. Based on the kinetic model constructed using the current temperature, the measured concentration, and the target desorption parameters, the signal deviation caused by the retention of gas molecules on the inner wall of the detection chamber is determined. The concentration of the gas to be measured is obtained by correcting the deviation of the measured concentration of the gas to be measured based on the signal deviation.

16. An electronic device, characterized in that, include: The device includes a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the design method of the detection cavity according to any one of claims 1 to 6, or to perform the gas concentration measurement method of the cleaning endpoint detection device according to claim 15.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to execute the design method of the detection cavity according to any one of claims 1 to 6, or to execute the gas concentration measurement method of the cleaning endpoint detection device according to claim 15.

18. A computer program product, characterized in that, It includes computer instructions for causing a computer to execute the design method of the detection cavity according to any one of claims 1 to 6, or to execute the gas concentration measurement method of the cleaning endpoint detection device according to claim 15.

Images