Apparatus, method, and computer readable medium for detecting gas leaks in semiconductor devices

Abstract

Provided is an apparatus for detecting a gas leak in a semiconductor device, including: a receiving unit configured to receive a first signal representing a sound collected from a first position along a gas path of the semiconductor device; a leak determining unit configured to determine whether a leak exists in the gas path in the vicinity of the first position based on the first signal; and an alarm unit configured to issue an alarm in response to a determination that a leak exists in the gas path in the vicinity of the first position. Discomfort and subjectivity of a detector can be avoided in gas leak detection, and detection of a minute leak can be achieved.

Description

Apparatus, methods, and computer-readable media for detecting gas leaks in semiconductor devices. Technical Field

[0001] This invention relates to gas leak detection, and more specifically, to gas leak detection in gas paths of semiconductor devices. Background Technology

[0002] Semiconductor equipment such as molding machines, wire bonding machines, and cutting machines extensively utilize pneumatic mechanisms powered by cylinders. Therefore, these devices contain numerous pneumatic connection devices, such as air pumps and valves. Maintenance and repair of this equipment necessitates checking for leaks in the pneumatic circuits. Since the cylinders in these semiconductor devices must remain operational during testing, the pneumatic mechanisms pose a potential hazard to personnel. Therefore, during testing, the semiconductor equipment must be shielded with a transparent cover.

[0003] Typically, leak detectors need to bend over and listen to the sound of leaks through a transparent enclosure. Maintaining this posture for extended periods can cause fatigue and injury. Furthermore, the transparent enclosure can obstruct human hearing, and the varying sensitivity of individual ears makes this judgment subjective, thus compromising accuracy. Alternatively, a straight rod with a ribbon wrapped around it can be inserted into the transparent enclosure to detect leaks. However, in this method, the rod may struggle to deliver the ribbon to every location being tested, and some minor leaks may go undetected due to insufficient ribbon movement.

[0004] Therefore, for semiconductor equipment, there is still a need to provide improved gas leak detection. Summary of the Invention

[0005] This invention provides an improved method for detecting gas leaks in semiconductor devices, which avoids discomfort and subjectivity for the detection personnel and enables the detection of minute leaks.

[0006] According to one aspect of the present invention, an apparatus for detecting gas leaks in a semiconductor device is provided, comprising: a receiving unit for receiving a first signal representing sound collected from a first location along a gas path of the semiconductor device; a leak determination unit for determining, based on the first signal, whether a leak exists in the gas path near the first location; and an alarm unit for issuing an alarm in response to determining that a leak exists in the gas path near the first location.

[0007] According to another aspect of the present invention, a method for detecting gas leaks in a semiconductor device is provided, comprising: receiving a first signal representing sound collected from a first location along a gas path of the semiconductor device; determining, based on the first signal, whether a leak exists in the gas path near the first location; and issuing an alarm in response to determining that a leak exists in the gas path near the first location.

[0008] According to another aspect of the invention, a computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform a method according to any embodiment of the invention.

[0009] In semiconductor devices, the sound near the gas path differs depending on whether a leak is present. Current acoustic sensors can acquire sound with high sensitivity across a wide frequency range, allowing the signals representing the acquired sound to reveal minute differences. Therefore, by analyzing the signals representing the acquired sound, these minute differences can be identified, enabling the detection of normally undetectable micro-leakage. This signal analysis can be automated by computing devices such as computers, eliminating the need for human judgment and simplifying the process while avoiding subjective errors inherent in human hearing.

[0010] Those skilled in the art will recognize other advantages of the invention after reading and understanding the following detailed description. Attached Figure Description

[0011] Figure 1 shows a block diagram of an apparatus 100 for detecting gas leaks in a semiconductor device according to an embodiment of the present invention;

[0012] Figure 2 shows a signal diagram 200 according to an embodiment of the present invention; and

[0013] Figure 3 shows a flowchart of a method 300 for detecting gas leakage in a semiconductor device according to an embodiment of the present invention. Detailed Implementation

[0014] Figure 1 shows a block diagram of an apparatus 100 for detecting gas leaks in a semiconductor device according to an embodiment of the present invention.

[0015] The device 100 includes a receiving unit 110 for receiving a first signal representing sound collected from a first location along the gas path of the semiconductor device.

[0016] An acoustic sensor 10, such as a microphone, can be used to acquire sound at a first location along the air path. For example, the acoustic sensor 10 can be a microelectromechanical system (MEMS) microphone, which has advantages such as miniaturization and stable performance. Such advantages facilitate delivering the microphone to different locations along the air path to accurately acquire sound at the corresponding locations. The microphone can, for example, generate a digital signal representing the acquired sound as a first signal using an analog-to-digital converter (A / D).

[0017] After the first signal is generated, a computing device such as a computer can receive the digital signal in its memory or cache. An application program in the computer can receive the digital signal for subsequent analysis. Therefore, the receiving unit 110 can represent a hardware receiving component within a device such as a computer, or it can represent a software module in an application program for receiving signals.

[0018] In one embodiment, the device 100 does not include a sound sensor 10, and the sound sensor 10 can be any sound sensor that can be obtained by other means, for example, the sound sensor 10 can be a microphone as an external device of the computer.

[0019] In another embodiment, the sound sensor 10 may be included in the device 100 as part of the device 100.

[0020] The primary location for sound collection can be near connecting devices in the gas path, such as near the air pump or valve along the gas path. These connecting devices are typically the components where leaks most frequently occur. For example, a gas leak may occur if the air pump or valve is not tightened. Therefore, it is necessary to collect sound near connecting devices such as air pumps and valves for use in leak detection.

[0021] Alternatively, the first location could be another point along the gas pipeline. Damage to the pipeline wall is also a potential cause of gas leaks. Therefore, it may also be necessary to collect sound samples near the pipeline wall for use in leak detection.

[0022] In the actual maintenance or repair of semiconductor equipment, leak detection is required at multiple locations, such as each connection device and each pipe wall, which need to be inspected sequentially. Therefore, sound needs to be collected at different locations. A sound sensor 10 can be attached to a rod 20 to facilitate delivery of the sound sensor 10 to different locations, thereby sequentially collecting sound at each location. For example, the rod 20 can be held by a technician, who can manually deliver the sound sensor 10 to different locations. Alternatively, the rod 20 can be held by a robotic arm, and the controlled movement of the robotic arm enables automated sound collection at multiple locations.

[0023] Preferably, the rod 20 is a bendable rod, so that the rod 20 can change between a straight state and a bend state with various bending shapes.

[0024] For example, the flexible rod 20 can be made from a soft metal such as aluminum. These soft metals allow for the rod to be bent and shaped, enabling the inspector to easily adjust the flexible rod 20 to various desired shapes. This is advantageous when it is necessary to extend the sound sensor 10 through the rod 20 into the transparent enclosure. The flexible shape facilitates the delivery of the sound sensor 10 to various locations within the transparent enclosure.

[0025] Alternatively, the bendable rod 20 can be formed using multiple joints. For example, the bending deformation of the bendable rod 20 can be achieved using hinge joints. The bendable rod 20 can also be deformed in three degrees of freedom using ball joints, thereby enabling the sound sensor 10 to be delivered to various locations with greater flexibility.

[0026] In one embodiment, the device 100 does not include the rod 20. The inspector can use any rod-shaped object capable of securing the sound sensor 10 (e.g., a screwdriver) as the rod 20.

[0027] In another embodiment, the rod 20 may be included in the device 100 as part of the device 100.

[0028] Instead of using rod 20 to deliver acoustic sensor 10 to different locations, acoustic sensor 10 may also include multiple sub-sensors, which are respectively arranged at multiple locations where leaks need to be detected. For example, multiple sub-sensors can be arranged near various connection devices and pipe walls. Each sub-sensor can be uniquely numbered, with each number corresponding to a specific location. For example, sub-sensor 1 is arranged near the air pump, sub-sensor 2 is arranged near the air valve, and sensor 3 is arranged near the pipe wall. Sound is collected from the air pump, air valve, and pipe wall respectively through these three sub-sensors.

[0029] In one embodiment, the first signal received by the receiving unit 110 may be the original sound wave signal of the acquired sound. The original sound wave signal represents the sound intensity that changes over time.

[0030] In another embodiment, the first signal received by the receiving unit 110 may be a pitch signal of the acquired sound. Each signal value of the pitch signal represents the energy of the sound acquired within a predetermined time period (e.g., 1 second).

[0031] For example, the energy of sound collected within a predetermined time period of 1 second can be obtained by calculating the root mean square of the original sound wave signal within that 1-second time period. This energy value can be used as the amplitude of the pitch signal for that 1-second time period. By calculating the energy value for consecutive 1-second time periods, the pitch signal that varies with time can be obtained.

[0032] For example, the energy of sound collected within a predetermined 1-second time period can be obtained by calculating the root mean square of the spectral signal. This energy value can be used as the amplitude of the tone signal for that 1-second time period. The spectral signal can be obtained by performing a Fourier transform on the original sound wave signal within that 1-second time period. Similarly, by calculating the energy values ​​for consecutive 1-second time periods, the tone signal that varies over time can be obtained.

[0033] The device 100 also includes a leak determination unit 120, which is used to determine whether there is a leak in the gas path near the first location based on the received first signal.

[0034] The leak detection unit 120 can determine whether a leak exists by detecting whether the amplitude of a first signal exceeds an amplitude threshold. When a leak occurs, an airflow sound is generated, so the sound intensity is expected to be greater when a leak occurs than when no leak occurs. The greater sound intensity is reflected in a larger amplitude of the sound wave signal or tone signal. Therefore, when the amplitude of the first signal exceeds a specific amplitude threshold, it indicates the presence of a leak. The specific amplitude threshold can be determined based on the experience of a professional. This threshold can also be continuously optimized through actual measurements obtained under both detected and undetected leak conditions.

[0035] Of course, a leak cannot be determined solely by the first signal briefly exceeding the threshold. If, during leak detection, an object suddenly falls, the resulting sound spike might cause the first signal to suddenly exceed the threshold; this should not be considered a leak. Therefore, a leak can only be confirmed by continuously detecting the signal exceeding the threshold within a predetermined time window (e.g., 20 seconds). Furthermore, since sound waves or tone signals typically exhibit oscillating waveforms, it is difficult to detect a signal that consistently exceeds the threshold within a predetermined time window. If a leak exists, the oscillating sound wave or tone signal may manifest as repeatedly exceeding the threshold. Therefore, a leak can only be confirmed by detecting the first signal exceeding the threshold a first predetermined number of times within the first predetermined time window.

[0036] In this embodiment, using a tone signal is advantageous because it significantly reduces the number of threshold comparisons. For example, with a microphone sampling frequency of 44kHz, the original acoustic signal has 44,000 signal values ​​over a 1-second time period. Taking a 20-second time window as an example, this would generate 880,000 signal values. The number of tone signal values ​​(as described above, representing the energy of sound within a predetermined time period, such as 1 second) can be drastically reduced. For instance, in an embodiment that calculates energy values ​​for each 1-second time period, the 44,000 original acoustic signal values ​​within 1 second can be converted into one tone signal value for that 1-second time period. For a 20-second time window, 20 tone signal values ​​can be generated. Therefore, for this 20-second time window, using the original acoustic signal requires 880,000 threshold comparisons; while using the tone signal, only 20 threshold comparisons are needed, which significantly reduces computational resources.

[0037] Ideally, the amplitude threshold representing the boundary between leakage and non-leakage should be consistent. However, semiconductor devices themselves generate ambient noise during operation. For example, as mentioned above, testing semiconductor devices typically requires keeping the cylinders running, which actuates the pneumatic mechanisms and produces sound. Different semiconductor devices may produce different ambient noise levels. For instance, semiconductor devices with more pneumatic mechanisms will produce more ambient noise compared to those with fewer.

[0038] When a microphone captures sound at the first location, in addition to potential sound leakage, it may also pick up ambient sounds generated by the operation of the semiconductor device. Since different semiconductor devices produce different ambient sounds, a uniform leakage threshold cannot be applied to all devices. Furthermore, in some cases, the amplitude of the ambient sound signal alone may exceed the amplitude threshold. Therefore, it is necessary to consider ambient sound and adjust the appropriate amplitude threshold for different semiconductor devices.

[0039] In one embodiment, in addition to the first signal, the receiving unit 110 also receives an environmental signal representing ambient sound, and the environmental signal is of the same type as the first signal. For example, if the first signal is a raw sound wave signal, then the environmental signal is also a raw sound wave signal; and if the first signal is a tone signal, then the environmental signal is also a tone signal.

[0040] In this embodiment, the leakage determination unit 120 can also adjust the amplitude threshold based on the environmental signal. For example, the maximum amplitude (or average amplitude) of the environmental signal can be added to the original threshold to form the adjusted threshold.

[0041] Since the ambient noise generated by the operation of a specific semiconductor device is generally stable and predictable, a threshold adjustment can be performed only once for that device. This adjusted threshold can be recorded and directly recalled in subsequent inspections of that type of semiconductor device.

[0042] In addition to the ambient noise generated by the stable operation of semiconductor equipment, unexpected continuous sounds such as a telephone ringing may suddenly occur during leak detection. Unlike the ambient noise generated by the stable operation of semiconductor equipment, these unexpected ambient noises are unpredictable. Therefore, it is difficult to compensate for such noises by adjusting thresholds.

[0043] To address the problem caused by such unexpected ambient sounds, in another embodiment, the receiving unit 101 also receives an ambient signal representing the ambient sound, and this ambient signal has the same type as the first signal. For example, if the first signal is a raw sound wave signal, then the ambient signal is also a raw sound wave signal; and if the first signal is a tone signal, then the ambient signal is also a tone signal.

[0044] In this embodiment, the leakage determination unit 120 first compensates for the first signal. As explained above, the sound collected at the first location may include not only sounds caused by possible leakage, but also unexpected ambient sounds such as telephone ringing. In this case, the first signal represents the sum of both the leakage sound and the ambient sound. Therefore, the first signal needs to be compensated in order to obtain a compensated first signal that represents only the leakage sound.

[0045] For example, the ambient signal can be subtracted from the first signal to filter out the ambient sound represented by the ambient signal from the total sound collected by the microphone.

[0046] Ambient sounds can be acquired by a second acoustic sensor to generate an ambient signal. This second acoustic sensor may be, for example, another microphone, which may be located outside a transparent enclosure used in semiconductor device detection to acquire unexpected ambient sounds such as telephone rings or conversations. Optionally, the second acoustic sensor may be included in device 100 to form part of device 100. Alternatively, the second acoustic sensor may be a component independent of device 100. For example, the second acoustic sensor may be the system microphone of the computing device used.

[0047] It is possible to consider using the attenuation effect of the transparent enclosure to generate an environmental signal when external ambient sound is transmitted into the enclosure, so that the environmental signal can more accurately reflect the intensity of the external ambient sound when it is collected at the first location inside the transparent enclosure.

[0048] Subsequently, the leak determination unit 120 can compare the compensated first signal with an amplitude threshold to determine whether there is a leak in the gas path near the first location, as described in the embodiments above.

[0049] Sometimes, there exist sounds that are neither recognized as sounds produced by semiconductor devices during stable operation nor captured by microphones that collect ambient sound for compensation. For example, inside a transparent enclosure, there might be fault sounds caused by temporary malfunctions in the pneumatic mechanism (such as friction). This sound can cause the sound signal to exceed the threshold even in locations where there is no leakage.

[0050] Therefore, after determining that the amplitude of the first signal at the first location exceeds the threshold, it is also necessary to determine whether the amplitude of the second signal at other locations also exceeds the threshold. The second signal represents the sound collected at locations other than the first location. The sound collected at other locations includes ambient sounds that could also occur at the first location, such as the aforementioned fault sound. Unlike the sound collected at the first location, the sound collected at other locations does not include the possible leakage sound at the first location. Therefore, if the amplitude of the second signal obtained at other locations, for example, even at locations far from the entire gas path, still exceeds the threshold, it is very likely that the cause of the signal exceeding the threshold is not a gas path leak, but rather ambient sounds such as a fault sound. In this case, even if the amplitude of the first signal at the first location exceeds the threshold, it cannot be completely determined whether there is a gas path leak near the first location, and further detection is required.

[0051] Only when the amplitude of the first signal at the first location exceeds the threshold, while the amplitude of the second signal at the second location does not exceed the threshold, can it be ensured that the amplitude of the first signal at the first location exceeding the threshold is due to gas leakage.

[0052] Therefore, in yet another embodiment, in addition to the first signal, the receiving unit 110 also receives a second signal representing sound collected from a second location away from the first location, the second signal having the same type as the first signal. For example, if the first signal is a raw sound wave signal, then the second signal is also a raw sound wave signal; and if the first signal is a tone signal, then the second signal is also a tone signal.

[0053] The reception of the first signal and the reception of the second signal can be completed simultaneously or sequentially. For example, the first and second signals can be acquired simultaneously by sound sensors distributed at different locations, allowing them to be received at the same time. Alternatively, the sound sensor can be first moved to a first location to acquire the first signal, thus completing the reception of the first signal; then the sound sensor can be moved to a second location to acquire the second signal, and then the reception of the second signal can be completed.

[0054] For example, the lever 20 described above can be used to move the sound sensor from the first position to the second position.

[0055] In this embodiment, the leak determination unit 120 detects whether the amplitudes of the first signal and the second signal exceed a threshold. For example, it can detect whether the amplitude of the first signal exceeds the amplitude threshold a first predetermined number of times within a first predetermined time window, and it can detect whether the amplitude of the second signal exceeds the amplitude threshold a second predetermined number of times within a second predetermined time window. Based on the fact that the amplitude of the first signal exceeds the amplitude threshold a first predetermined number of times within the first predetermined time window, and the amplitude of the second signal does not exceed the amplitude threshold a second predetermined number of times within the second predetermined time window, the leak determination unit 120 determines that there is a leak in the gas path near the first location.

[0056] The first predetermined time window can be the same as or different from the second predetermined time window. For example, if it is determined that the amplitude of the first signal exceeds the threshold within 20 seconds, it is only necessary to determine that the amplitude of the second signal does not exceed the threshold within 2 seconds to confirm that no other sounds are affecting the second position, thereby determining that there is a gas leak near the first position. Similarly, the first predetermined number of times can be the same as or different from the second predetermined number of times.

[0057] Instead of the threshold method described above, the leakage determination unit 120 can also determine whether a leakage has occurred using a machine learning algorithm. This machine learning algorithm trains a machine learning model using information such as the specific type of the semiconductor device, instances of leakage, and the first signal corresponding to those instances. This allows the machine learning model to be used to determine whether a leakage exists based on the actually acquired first signal.

[0058] The leakage determination unit 120 may represent a hardware component such as a computer processor or an application-specific integrated circuit, or it may represent a software module in an application used to determine leakage.

[0059] The device 100 also includes an alarm unit 130. After the leak detection unit 120 determines that a leak exists, the alarm unit 130 can issue an alarm to alert the inspection personnel to the leak.

[0060] Alarms can be visual, auditory, tactile, or any other conceivable type. For example, after a leak is detected, the inspector may be alerted by flashing a computer screen or any other indicator light; or by an alarm sound emanating from a speaker; or by vibration from a handheld computing device (such as a mobile phone or tablet).

[0061] In another embodiment, the inspectors may not be present at the inspection site. Upon detection of a leak, an alert can be sent to the inspectors via email or SMS to remind them to complete subsequent maintenance.

[0062] Of course, an alarm can be any combination of the various types of alarms mentioned above.

[0063] Alarm unit 130 can refer to hardware such as a light or a speaker used to issue an alarm, or it can refer to a software module in an application used to control the corresponding hardware to issue an alarm.

[0064] Optionally, the device 100 also includes a display unit 140, which can display the acoustic waveform of the original acoustic signal or the pitch waveform of the pitch signal on a display for inspection personnel to view.

[0065] After an alarm is triggered, inspectors can more intuitively confirm whether a leak has occurred by viewing a sound wave diagram or tone diagram.

[0066] Preferably, the threshold can be displayed together with the acoustic waveform or tone graph. For example, the threshold can be shown as a line in the acoustic waveform or tone graph, allowing the inspector to visually observe whether the signal amplitude in the acoustic waveform or tone graph exceeds the line representing the threshold multiple times within a first predetermined time window, thereby more intuitively confirming whether a leak has occurred.

[0067] The display unit 140 may represent a display device such as a computer monitor or a tablet computer touch screen, or it may represent a software module in an application for controlling the display of graphics on the corresponding hardware.

[0068] Figure 2 illustrates a signal diagram 200 according to an embodiment of the present invention, which can help inspectors visually determine whether a leak has occurred near a first location. In this embodiment, the signal diagram 200 is a tone diagram. In this tone diagram, the horizontal axis represents time in seconds (s), and the vertical axis represents the amplitude value of the tone signal.

[0069] In this graph, the vertical axis shows the amplitude range from -4000 to 4000. This value has no actual unit, and users, such as testing personnel, can select any suitable range for easy observation. The measured signal value can be mapped to this range. For example, the minimum measured signal value can be mapped to -4000 and the maximum measured signal value can be mapped to 4000, with intermediate signal values ​​linearly mapped between -4000 and 4000.

[0070] In this embodiment, 2000 was selected as the amplitude threshold. It can be seen that during the time window of 350s-370s, the microphone is delivered to the first position along the air path. During this time, the signal amplitude repeatedly exceeds the 2000 threshold. After 370s, the microphone is moved to a second position away from the first position, at which point the signal amplitude no longer exceeds the 2000 threshold. Therefore, it can be determined that there is a leak in the air path near the first position.

[0071] Figure 3 shows a flowchart of a method 300 for detecting gas leakage in a semiconductor device according to an embodiment of the present invention.

[0072] In step 310, a first signal representing sound collected from a first location along the gas path of the semiconductor device is received.

[0073] In step 320, it is determined whether there is a leak in the gas path near the first location based on the first signal.

[0074] As described in detail in the embodiments of the reference device above, a leak in the gas path near the first location can be determined by detecting that the amplitude of the first signal exceeds a threshold a first predetermined number of times within a first predetermined time window.

[0075] Alternatively, as described in detail in another embodiment of the reference device above, the first signal can be used as input to a machine learning algorithm to determine that a leak exists in the gas path near the first location.

[0076] In step 330, an alarm is issued in response to determining that a leak exists in the gas path near the first location.

[0077] It is understood that the method according to the invention has the same or similar embodiments as the apparatus according to the invention.

[0078] The method shown in Figure 3 can be implemented by a processor executing corresponding instructions. These instructions can be stored on any suitable computer-readable medium.

[0079] The method of the present invention has been described above with reference to the embodiment shown in FIG3 only. It is understood that the various operations included in the above embodiments are not limiting and can be deleted, combined, changed, split and / or recombined as needed to add / modify / delete the corresponding functions.

[0080] The apparatus and method of the present invention have been described above with reference to various embodiments, wherein the embodiments mentioned may include specific features, structures, or characteristics, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, some embodiments may have some or all of the features described in other embodiments, or may not have the features described in other embodiments.

[0081] Various features of different embodiments or examples can be combined in various ways with some included features and others excluded features to adapt to a variety of different applications. The accompanying drawings and the foregoing description provide examples of embodiments. Those skilled in the art will understand that one or more of the described elements can be combined into a single functional element. Alternatively, certain elements can be divided into multiple functional elements. Elements from one embodiment can be added to another embodiment. For example, the order of processes described herein can be changed and is not limited to the manner described herein. Furthermore, the operations of any flowchart need not be performed in the order shown; nor is it necessary to perform all operations. Moreover, those operations that do not depend on other operations can be performed in parallel with other operations. The scope of the embodiments is by no means limited to these specific examples. Many variations, such as differences in the order of operations, product composition, and structure, are possible, whether or not explicitly stated in the specification.

Claims

1. An apparatus for detecting gas leaks in semiconductor equipment, comprising: A receiving unit is configured to receive a first signal representing sound collected from a first location along the gas path of the semiconductor device and a second signal representing sound collected from a second location away from the first location, the second signal having the same type as the first signal; a leakage determination unit is configured to: detect whether the amplitude of the first signal exceeds an amplitude threshold a first predetermined number of times within a first predetermined time window, detect whether the amplitude of the second signal exceeds the amplitude threshold a second predetermined number of times within a second predetermined time window, and in response to detecting that the amplitude of the first signal exceeds the amplitude threshold a first predetermined number of times within the first predetermined time window and the amplitude of the second signal does not exceed the amplitude threshold a second predetermined number of times within the second predetermined time window, determine that there is a leakage in the gas path near the first location; And an alarm unit for issuing an alarm in response to determining that a gas leak exists in the vicinity of the first location.

2. The apparatus according to claim 1, wherein, The receiving unit is further configured to receive an environmental signal representing ambient sound, the environmental signal having the same type as the first signal, wherein the leakage determination unit is further configured to: adjust the amplitude threshold based on the environmental signal.

3. The apparatus according to claim 1, wherein, The receiving unit is further configured to receive an environmental signal representing ambient sound, the environmental signal having the same type as the first signal, wherein the leakage determination unit is configured to: compensate the first signal with the environmental signal; detect whether the amplitude of the compensated first signal exceeds an amplitude threshold a first predetermined number of times within a first predetermined time window; and, in response to detecting that the amplitude of the compensated first signal exceeds the amplitude threshold a first predetermined number of times within the first predetermined time window, determine that there is a leak in the gas path near the first location.

4. The apparatus according to any one of claims 1-3, wherein, The first signal is a raw acoustic wave signal, and the device further includes a display unit for displaying an acoustic waveform of the raw acoustic wave signal.

5. The apparatus according to any one of claims 1-3, wherein, The first signal is a tone signal, each signal value of the tone signal represents the energy of the sound collected within a predetermined time period, and the device further includes a display unit for displaying a tone graph of the tone signal.

6. The apparatus according to any one of claims 1-3, further comprising: An acoustic sensor for acquiring sound from the first location along the gas path of the semiconductor device to generate the first signal; And a flexible rod, to which the acoustic sensor is attached so that the acoustic sensor can be delivered to different locations.

7. The apparatus according to any one of claims 1-3, further comprising: An acoustic sensor comprising a plurality of sub-sensors arranged at different locations, wherein one of the plurality of sub-sensors is used to acquire sound from the first location along the gas path of the semiconductor device to generate the first signal.

8. The apparatus according to claim 3, further comprising: A second acoustic sensor is used to collect ambient sounds to generate the ambient signal.

9. A method for detecting gas leakage in a semiconductor device, comprising: The system receives a first signal representing sound collected from a first location along the gas path of the semiconductor device and a second signal representing sound collected from a second location away from the first location, the second signal being of the same type as the first signal; detects whether the amplitude of the first signal exceeds an amplitude threshold a first predetermined number of times within a first predetermined time window, detects whether the amplitude of the second signal exceeds the amplitude threshold a second predetermined number of times within a second predetermined time window, and determines that there is a leak in the gas path near the first location in response to detecting that the amplitude of the first signal exceeds the amplitude threshold a first predetermined number of times within the first predetermined time window and the amplitude of the second signal does not exceed the amplitude threshold a second predetermined number of times within the second predetermined time window; and issues an alarm in response to determining that there is a leak in the gas path near the first location.

10. The method of claim 9, further comprising: Display the signal diagram of the first signal.

11. A computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform the method according to any one of claims 9-10.

Images