TW201502302A - Sputtering process control system capable of minute adjustment of process parameters and method thereof - Google Patents
Sputtering process control system capable of minute adjustment of process parameters and method thereof Download PDFInfo
- Publication number
- TW201502302A TW201502302A TW102125164A TW102125164A TW201502302A TW 201502302 A TW201502302 A TW 201502302A TW 102125164 A TW102125164 A TW 102125164A TW 102125164 A TW102125164 A TW 102125164A TW 201502302 A TW201502302 A TW 201502302A
- Authority
- TW
- Taiwan
- Prior art keywords
- control system
- sputtering
- parameter
- neurons
- control
- Prior art date
Links
Landscapes
- Physical Vapour Deposition (AREA)
Abstract
Description
本發明係有關於一種濺鍍製程控制系統及其方法，其尤指一種利用類神經網路模糊控制法則微調整製程參數並無須建置資料庫的濺鍍製程控制系統及其方法。
The invention relates to a sputtering process control system and a method thereof, in particular to a sputtering process control system and a method thereof for using the neural network fuzzy control rule to finely adjust process parameters without constructing a database.
從製程技術的觀點來說，先進製程控制可以說是全部製程的整合者，尤其是在每層薄膜對成品品質影響巨大的結構之中，應用先進製程控制能夠達到精準地監控與調校製程參數。From the point of view of process technology, advanced process control can be said to be the integrator of all processes, especially in the structure where each layer of film has a great influence on the quality of finished products. Advanced process control can be used to accurately monitor and adjust process parameters. .
請參閱第一圖，目前濺鍍製程控制系統1’主要包含一控制單元12’、一資料庫13’及至少一檢測器10’，控制單元12’及檢測器10’連接一製程設備2’，資料庫13’連接控制單元12’。資料庫13’主要用來記錄製程之各項資料，以提供控制策略進行運算所需之歷史資料。控制單元12’依據控制策略來對取得之資料進行運算，因此計算與讀取資料之速度會影響濺鍍製程控制系統1’之反應時間。Referring to the first figure, the current sputtering process control system 1' mainly includes a control unit 12', a database 13' and at least one detector 10'. The control unit 12' and the detector 10' are connected to a process device 2'. The database 13' is connected to the control unit 12'. The database 13' is mainly used to record various data of the process to provide historical data required for the control strategy to perform operations. The control unit 12' operates on the acquired data in accordance with the control strategy, so that the speed at which the data is read and read affects the reaction time of the sputtering process control system 1'.
顯然目前之製程控制系統係以資料庫之架構進行製程控制，有以精簡計算方程式之方法來減低處理器負擔，亦有加強硬體效能以雙核運算概念來加速製程控制系統反應之方法，但均以具有資料庫之架構來進行整體之控制。然而隨著鍍膜製程發展，越來越多製程開始出現大量之參數資料，製程品質相關性之運算也越加複雜，造成資料庫之建置也造成設備成本與耗材成本墊高，嚴重影響量產之成本。Obviously, the current process control system is based on the structure of the database for process control. It has the method of streamlining the calculation equation to reduce the processor load, and also has the method of strengthening the hardware performance to accelerate the process control system response with the concept of dual-core computing. The overall control is based on the architecture of the database. However, with the development of coating process, more and more processes begin to appear a large number of parameter data, and the calculation of process quality correlation is more complicated. As a result, the construction of the database also causes the cost of equipment and consumables to rise, which seriously affects mass production. The cost.
有鑑於上述問題，本發明提供一種可微調整製程參數之濺鍍製程控制系統及其方法，本發明之濺鍍製程控制系統未額外建置資料庫，並採用具有學習能力之類神經模糊控制法則對製程參數進行微調整，如此有效改善上述問題。
In view of the above problems, the present invention provides a sputtering process control system and a method thereof for micro-adjusting process parameters. The sputtering process control system of the present invention does not additionally build a database, and adopts a neural fuzzy control rule with learning ability. Fine adjustment of the process parameters is effective in improving the above problems.
本發明之目的，係提供一種可微調整製程參數之濺鍍製程控制系統及其方法，其未額外設有資料庫，因本發明利用具有學習能力之類神經模糊控制法則微調整製程參數，如此節省建構設備之成本，並節省對資料庫進行搜尋與資料統計運算之時間，進而加快控制系統之反應時間與參數調校之效率。The object of the present invention is to provide a sputtering process control system and method thereof for micro-adjusting process parameters, which are not additionally provided with a database, because the present invention uses a neural fuzzy control rule with learning ability to finely adjust process parameters, It saves the cost of constructing equipment and saves time for searching and data statistics of the database, thereby accelerating the response time of the control system and the efficiency of parameter adjustment.
為了達到上述所指稱之各目的與功效，本發明係揭示了一種可微調整製程參數之濺鍍製程控制系統，其包含：至少一檢測單元，其連接一製程設備，並檢測該製程設備以產生一檢測資訊；以及一控制單元，其連接該製程設備及該檢測單元，該控制單元接收至少一時間參數、至少一製程參數及該檢測單元所提供之該檢測資訊，並依據一類神經模糊控制法則微調整該製程參數，以產生一控制命令，該控制單元傳輸該控制命令至該製程設備，該製程設備依據該控制命令執行。In order to achieve the above-mentioned various purposes and effects, the present invention discloses a sputtering process control system capable of micro-adjusting process parameters, comprising: at least one detecting unit connected to a process device, and detecting the process device to generate a detection unit; and a control unit connected to the process device and the detection unit, the control unit receiving at least one time parameter, at least one process parameter, and the detection information provided by the detection unit, and according to a class of neural fuzzy control rules The process parameter is finely adjusted to generate a control command, and the control unit transmits the control command to the process device, and the process device executes according to the control command.
本發明揭示了一種可微調整製程參數之濺鍍製程控制方法，其包含：匯入一時間參數、至少一製程參數及一檢測資訊至一濺鍍製程控制系統；該濺鍍製程控制系統依據一類神經模糊控制法則運算該時間參數、該製程參數及該檢測資訊，以微調整該製程參數，並依據微調整之該製程參數產生一控制命令；該濺鍍製程控制系統傳輸該控制命令至一製程設備；以及該製程設備依據該控制命令進行濺鍍製程。
The invention discloses a sputtering process control method capable of micro-adjusting process parameters, comprising: importing a time parameter, at least one process parameter and a detection information to a sputtering process control system; the sputtering process control system is according to a class The neural fuzzy control rule calculates the time parameter, the process parameter and the detection information to finely adjust the process parameter, and generates a control command according to the micro-adjusted process parameter; the sputtering process control system transmits the control command to a process And the process device performs a sputtering process according to the control command.
第一圖：其為習知之濺鍍製程控制系統的示意圖；
第二圖：其為本發明之第一實施例之濺鍍製程控制系統的示意圖；
第三圖：其為本發明之第一實施例之濺鍍製程控制方法的流程圖；以及
第四圖：其為本發明之第一實施例之類神經模糊控制法則的示意圖。
First: it is a schematic diagram of a conventional sputtering process control system;
Second drawing: a schematic view of a sputtering process control system according to a first embodiment of the present invention;
Third Embodiment: A flowchart of a sputtering process control method according to a first embodiment of the present invention; and a fourth diagram: a schematic diagram of a neuro-fuzzy control law according to a first embodiment of the present invention.
為使對本發明之特徵及所達成之功效有更進一步之瞭解與認識，謹佐以實施例及配合詳細之說明與圖式，說明如後：In order to further understand and understand the features of the present invention and the effects achieved, the embodiments and the detailed description and drawings are described as follows:
習知濺鍍製程控制系統均建置資料庫進行製程參數調整，因此增加了對資料庫進行搜尋與資料統計運算之時間，導致製程控制系統之反應時間增加。另外因目前製程越來越複雜，並開始產生大量之參數資料，而且製程品質相關性之運算也越加複雜，因此造成資料庫之資料量龐大、控制系統負擔過大，也會影響製程控制系統之反應時間，同時資料庫之建置也造成設備成本提升，嚴重影響量產之成本。Conventional sputter control system has built a database to adjust the process parameters, thus increasing the time for searching and data statistics of the database, resulting in an increase in the reaction time of the process control system. In addition, because the current process is more and more complicated, and a large amount of parameter data is generated, and the calculation of the process quality correlation is more complicated, the data volume of the database is huge, the control system is overburdened, and the process control system is also affected. The reaction time and the establishment of the database also increase the cost of equipment, which seriously affects the cost of mass production.
有鑒於上述問題，本發明提供一種可微調整製程參數之濺鍍製程控制系統及其方法，其利用具有學習能力之類神經模糊控制法則微調整製程參數，而且無須額外建置資料庫，即無須對資料庫進行搜尋與資料統計運算，不但節省建置資料庫之成本，更節省對資料庫進行搜尋與資料統計運算之時間，有效提高濺鍍製程控制系統之反應速度。In view of the above problems, the present invention provides a sputtering process control system and a method thereof for micro-adjusting process parameters, which use a neural fuzzy control rule with learning ability to finely adjust process parameters without additional database construction, that is, without Searching and data statistics on the database not only saves the cost of building the database, but also saves time for searching and statistical calculation of the database, effectively improving the response speed of the sputtering process control system.
請參閱第二圖，其為本發明之第一實施例之濺鍍製程控制系統的示意圖；如圖所示，本實施例提供一種濺鍍製程控制系統1，濺鍍製程控制系統1用於控制一製程設備2，本實施例之製程設備2可為應用於物理氣相沉積系統之直流濺鍍設備、交流濺鍍設備或射頻濺鍍設備。本實施例之濺鍍製程控制系統1包含至少一檢測單元10及一控制單元12。檢測單元10及控制單元12分別連接至製程設備2，控制單元12連接至檢測單元10，如此控制單元12輸入一控制命令至製程設備2，製程設備2依據控制命令進行濺鍍製程，然檢測單元10會檢測濺鍍製程中或後之狀態（如濺鍍製程之狀態和鍍膜之狀態），並產生至少一檢測資訊，且傳送檢測資訊至控制單元12。Please refer to the second figure, which is a schematic diagram of a sputtering process control system according to a first embodiment of the present invention; as shown, the embodiment provides a sputtering process control system 1 for controlling a sputtering process control system 1 A process device 2, the process device 2 of the embodiment may be a DC sputtering device, an AC sputtering device or an RF sputtering device applied to a physical vapor deposition system. The sputtering process control system 1 of the present embodiment includes at least one detecting unit 10 and a control unit 12. The detecting unit 10 and the control unit 12 are respectively connected to the processing device 2, and the control unit 12 is connected to the detecting unit 10, such that the control unit 12 inputs a control command to the processing device 2, and the processing device 2 performs a sputtering process according to the control command, and the detecting unit 10 will detect the state during or after the sputtering process (such as the state of the sputtering process and the state of the coating), generate at least one detection information, and transmit the detection information to the control unit 12.
控制單元12包含一處理器121，處理器121可為一微控制器、一數位信號處理器、一現場可規劃邏輯閘陣列或複雜可編程邏輯器件。控制單元12之控制命令的產生係複數參數及資訊經控制單元12內之處理器121運算而獲得，其中該些參數及資訊包含至少一時間參數、至少一製程參數及由檢測單元10所提供之檢測資訊，然時間參數及製程參數係可由一生產管理系統進行設定，並儲存於一記憶體（未繪示），其中時間參數為濺鍍製程實際進行之時間，其主要用於評估濺鍍製程控制系統1之參數漂移程度，以產生相對應之補償訊號；製程參數為濺鍍製程相關設定參數，生產管理系統所提供之製程參數為初始設定值，製程參數需經濺鍍製程控制系統1進行微調整；檢測資訊為濺鍍製程中或後之狀態，如電漿狀態、鍍膜品質，以得到誤差量值。The control unit 12 includes a processor 121, which may be a microcontroller, a digital signal processor, a field programmable logic gate array, or a complex programmable logic device. The generation of the control commands of the control unit 12 is obtained by the processor 121 in the control unit 12, wherein the parameters and information include at least one time parameter, at least one process parameter, and provided by the detecting unit 10. The detection information, the time parameter and the process parameter can be set by a production management system and stored in a memory (not shown), wherein the time parameter is the actual time of the sputtering process, which is mainly used to evaluate the sputtering process. The parameter drift degree of the control system 1 is generated to generate a corresponding compensation signal; the process parameters are setting parameters related to the sputtering process, the process parameters provided by the production management system are initial set values, and the process parameters are subjected to the sputtering process control system 1 Micro-adjustment; the detection information is the state in or after the sputtering process, such as the plasma state, the coating quality, to obtain the error magnitude.
請一併參閱第三圖，其為本發明之第一實施例之濺鍍製程控制方法的流程圖；如圖所示，本實施例之濺鍍製程控制方法係先執行步驟S10，匯入時間參數、製程參數及檢測單元10所提供之檢測資訊至控制單元12之處理器121，其中，該時間參數及製程參數係自該生產管理系統匯入控制單元12，而非自資料庫匯入。接著執行步驟S12，處理器121依據一類神經模糊控制法則運算時間參數、製程參數及檢測資訊，以微調整製程參數，並依據微調整後之製程參數產生控制命令。然後執行步驟S14，控制單元12之處理器121傳送控制命令至製程設備2，製程設備2依據控制命令進行濺鍍製程。接著執行步驟S16，檢測單元12檢測製程設備2進行濺鍍製程之一製程狀態，並依據該製程狀態產生更新之一檢測資訊，然後執行步驟S17，檢測單元12依據更新的檢測資訊判斷目前的製程狀態是否為理想之製程狀態，若濺鍍製程之狀態為理想之製程狀態，即可停止；若濺鍍製程之製程狀態未達到理想之製程狀態，則重覆執行步驟S10至步驟S17，直到檢測單元12檢測濺鍍製程之製程狀態為理想之製程狀態，即可停止。Please refer to the third figure, which is a flowchart of the sputtering process control method according to the first embodiment of the present invention; as shown in the figure, the sputtering process control method of the embodiment first performs step S10, and the import time is first. The parameter, the process parameter and the detection information provided by the detecting unit 10 are sent to the processor 121 of the control unit 12, wherein the time parameter and the process parameter are imported from the production management system into the control unit 12 instead of being imported from the database. Then, in step S12, the processor 121 calculates the time parameter, the process parameter and the detection information according to a type of neural fuzzy control rule to finely adjust the process parameter, and generates a control command according to the micro-adjusted process parameter. Then, in step S14, the processor 121 of the control unit 12 transmits a control command to the process device 2, and the process device 2 performs a sputtering process according to the control command. Next, in step S16, the detecting unit 12 detects that the process device 2 performs a process state of one of the sputtering processes, and generates one of the detection information according to the process state, and then performs step S17, and the detecting unit 12 determines the current process according to the updated detection information. Whether the state is an ideal process state, if the state of the sputtering process is an ideal process state, it can be stopped; if the process state of the sputtering process does not reach the ideal process state, step S10 to step S17 are repeatedly performed until the detecting unit 12 The process state of the sputtering process is determined to be the ideal process state, and can be stopped.
請一併參閱第四圖，其為本發明之第一實施例之類神經模糊控制法則的示意圖；如圖所示，本實施例之處理器121利用類神經模糊控制法則運算匯入至處理器121之時間參數、製程參數及檢測資訊，以微調整製程參數。類神經模糊控制法則具有一模糊化類神經網路架構3，模糊化類神經網路架構3為階層式網路，即由複數層構成，每一層具有複數神經元，每層之該些神經元不會相互連接，而每層之該些神經元會與上一層之該些神經元及下一層之該些神經元連接，然考量製程控制系統之反應時間，該些層之數量控制於3層與6層之間。然模糊化類神經網路架構3有多種型態，下述特舉出一種模糊化類神經網路架構3為例並進行說明。Please refer to the fourth figure, which is a schematic diagram of a neuro-fuzzy control rule according to the first embodiment of the present invention; as shown in the figure, the processor 121 of the present embodiment uses the neuro-fuzzy-like control algorithm to import the processor into the processor. 121 time parameters, process parameters and test information to fine tune process parameters. The neuro-fuzzy control law has a fuzzification-like neural network architecture 3. The fuzzification-like neural network architecture 3 is a hierarchical network, that is, composed of a plurality of layers, each layer having a plurality of neurons, and the neurons of each layer They are not connected to each other, and the neurons in each layer are connected to the neurons in the upper layer and the neurons in the next layer. However, the reaction time of the process control system is considered, and the number of the layers is controlled on the third layer. Between 6 floors. However, the fuzzification-like neural network architecture 3 has various types. A fuzzification-like neural network architecture 3 is exemplified and described below.
本實施例之模糊化類神經網路架構3具有五層，第一層31為輸入層，第一層31之該些神經元331之數量係依據匯入至處理器121之時間參數、製程參數及檢測資訊的數量而定，當然為了有較佳的製程控制系統1之反應時間，匯入至處理器121之時間參數、製程參數及檢測資訊的數量控制於2個與8個之間，進而保持製程控制系統1之運算效能，而輸入層31之每一神經元311的結構係與對應之參數進行歸屬函數的運算，即將參數映射於模糊集合中，也就是相容程度性之計算。本實施例之濺鍍製程控制方法係採用多參數輸入，有別於習知之製程控制方法僅單一參數輸入，如此有效提升調校製程參數之準確度，進而提升電漿品質及穩定性，更提高鍍膜品質。The fuzzification-like neural network architecture 3 of the present embodiment has five layers, and the first layer 31 is an input layer. The number of the neurons 331 of the first layer 31 is based on time parameters and process parameters that are imported into the processor 121. And the number of detection information, of course, in order to have a better reaction time of the process control system 1, the number of time parameters, process parameters, and detection information that are imported into the processor 121 is controlled between two and eight, and further The performance of the process control system 1 is maintained, and the structure of each neuron 311 of the input layer 31 performs the operation of the attribution function with the corresponding parameters, that is, the parameters are mapped into the fuzzy set, that is, the calculation of the degree of compatibility. The sputtering process control method of the embodiment adopts multi-parameter input, which is different from the conventional process control method, and only has a single parameter input, thereby effectively improving the accuracy of the calibration process parameters, thereby improving the quality and stability of the plasma, and improving the plasma. Coating quality.
第二層32之該些神經元321與輸入層31之該些神經元311相互連結，每一個連結具有一權重值，然後第二層32之每一神經元321的結果係依據其與輸入層31之該些神經元311之連結，找出對應之輸入層311之每一神經元311的結果與對應之權重值相乘積並進行「及」運算而獲得，第二層32之每一神經元321的結果即表示類神經模糊控制法則之前鑑部的啟動強度。The neurons 321 of the second layer 32 are interconnected with the neurons 311 of the input layer 31, each of the links having a weight value, and then the result of each of the neurons 321 of the second layer 32 is based on the input layer The connecting of the neurons 311 of 31, finding the result of each neuron 311 of the corresponding input layer 311 and multiplying the corresponding weight value and performing a "sum" operation, each nerve of the second layer 32 The result of element 321 represents the starting strength of the section before the neuro-fuzzy control law.
然第三層33之該些神經元331與第二層32之該些神經元321相互連結，每一個連結亦具有一權重值，而第三層33之每一神經元331的結果係依據其與第二層32之每一神經元321的連結，找出對應之第二層32之該些神經元321的結果及對應之該些連結的權重值進行正規化運算。The neurons 331 of the third layer 33 and the neurons 321 of the second layer 32 are connected to each other, and each of the links also has a weight value, and the result of each of the neurons 331 of the third layer 33 is based on the result. The connection with each of the neurons 321 of the second layer 32 finds the result of the neurons 321 corresponding to the second layer 32 and the weight values of the corresponding links are normalized.
接著第四層34之該些神經元341與第三層33之該些神經元331相互連結，每一個連結亦具有一權重值，而第四層34之每一神經元341的結果係依據其與第三層33之每一神經元331的連結，找出對應之第三層33之該些神經元331的結果、對應之該些連結的權重值與模糊模式進行運算，即進行類神經模糊控制法則之後鑑部的運算。Then, the neurons 341 of the fourth layer 34 and the neurons 331 of the third layer 33 are connected to each other, each of the links also has a weight value, and the result of each of the neurons 341 of the fourth layer 34 is based on Linking with each of the neurons 331 of the third layer 33, finding the result of the neurons 331 corresponding to the third layer 33, corresponding to the weight values of the links, and the fuzzy mode, that is, performing neuro-fuzzy Control the law after the law.
最後第五層35為輸出層，其僅具有單一神經元351，此神經元351分別與第四層34之該些神經元341連結，而每一連結亦具有一權重值，而此神經元351主要將第四層34之該些神經元341的結果進行加總，以產生經微調整之製程參數，並產生對應之控制命令。然檢測單元10檢測依據微調整之製程參數運作之製程設備2的狀態，以與理想的製程狀態進行比對運算，可得到製程設備2之狀態與理想的製程狀態間之誤差量值，以產生更新的檢測資訊，再輸入至控制單元12，以修正模糊化類神經網路架構3內之該些連結的權重值，所以本實施例之類神經模糊控制法則於每次執行後可微調整模糊化類神經網路架構3內之該些連結的權重值，以使微調整之製程參數達到理想的製程參數，如此也表示本實施例之類神經模糊控制法則本身具有學習能力，而不須透過外部建置之資料庫，所以可節省於資料庫中搜尋及統計運算資料的時間。Finally, the fifth layer 35 is an output layer having only a single neuron 351, which is respectively coupled to the neurons 341 of the fourth layer 34, and each link also has a weight value, and the neuron 351 The results of the neurons 341 of the fourth layer 34 are primarily summed to produce finely tuned process parameters and corresponding control commands are generated. The detecting unit 10 detects the state of the process device 2 operating according to the micro-adjusted process parameters, and performs a comparison operation with the ideal process state, so that the error magnitude between the state of the process device 2 and the ideal process state can be obtained to generate The updated detection information is input to the control unit 12 to correct the weight values of the links in the fuzzification-like neural network architecture 3. Therefore, the neural fuzzy control rule of the embodiment can finely adjust the blur after each execution. The weighting values of the links in the neural network architecture 3 are such that the process parameters of the micro-adjustment reach the desired process parameters, which also means that the neuro-fuzzy control law of the embodiment itself has the learning ability without Externally built database, so you can save time searching and counting data in the database.
綜上所述，本發明提供一種可微調整製程參數之濺鍍製程控制系統及其方法，其利用類神經模糊控制法則微調整製程參數，因類神經模糊控制法則有自學能力，而無須額外建置資料庫，有效降低建置設備之成本，並有效提升濺鍍製程控制系統之反應時間及調校製程參數之時間。本發明主要匯入多個參數至濺鍍製程控制系統，以有效調校製程參數之準確度，進而提升電漿品質及穩定性，更提升鍍膜品質。In summary, the present invention provides a sputtering process control system and method thereof for micro-adjusting process parameters, which use a neuro-fuzzy control law to finely adjust process parameters, and the neuro-fuzzy control law has self-learning ability without additional construction. The database is used to effectively reduce the cost of building equipment and effectively increase the reaction time of the sputtering process control system and adjust the process parameters. The invention mainly incorporates a plurality of parameters into the sputtering process control system to effectively adjust the accuracy of the process parameters, thereby improving the quality and stability of the plasma and improving the coating quality.
惟以上所述者，僅為本發明之較佳實施例而已，並非用來限定本發明實施之範圍，舉凡依本發明申請專利範圍所述之形狀、構造、特徵及精神所為之均等變化與修飾，均應包括於本發明之申請專利範圍內。
The above is only the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, and the variations, modifications, and modifications of the shapes, structures, features, and spirits described in the claims of the present invention. All should be included in the scope of the patent application of the present invention.
1’‧‧‧濺鍍製程控制系統
10’‧‧‧檢測器
12’‧‧‧控制單元
13’‧‧‧資料庫
2’‧‧‧製程設備
1‧‧‧濺鍍製程控制系統
10‧‧‧檢測單元
12‧‧‧控制單元
121‧‧‧處理器
2‧‧‧製程設備
3‧‧‧模糊化類神經網路架構
31‧‧‧第一層
311‧‧‧神經元
32‧‧‧第二層
321‧‧‧神經元
33‧‧‧第三層
331‧‧‧神經元
34‧‧‧第四層
341‧‧‧神經元
35‧‧‧第五層
351‧‧‧神經元1'‧‧‧ Sputtering Process Control System
10'‧‧‧Detector
12'‧‧‧Control unit
13'‧‧‧Database
2'‧‧‧Processing equipment
1‧‧‧ Sputtering Process Control System
10‧‧‧Detection unit
12‧‧‧Control unit
121‧‧‧ processor
2‧‧‧Processing equipment
3‧‧‧Fuzzy-like neural network architecture
31‧‧‧ first floor
311‧‧‧ neurons
32‧‧‧ second floor
321‧‧‧ neurons
33‧‧‧ third floor
331‧‧‧ neurons
34‧‧‧ fourth floor
341‧‧‧ neurons
35‧‧‧5th floor
351‧‧‧ neurons
1‧‧‧濺鍍製程控制系統 1‧‧‧ Sputtering Process Control System
10‧‧‧檢測單元 10‧‧‧Detection unit
12‧‧‧控制單元 12‧‧‧Control unit
121‧‧‧處理器 121‧‧‧ processor
2‧‧‧製程設備 2‧‧‧Processing equipment
Claims (11)
至少一檢測單元，其連接一製程設備，並檢測該製程設備以產生一檢測資訊；以及
一控制單元，其連接該製程設備及該檢測單元，該控制單元接收至少一時間參數、至少一製程參數及該檢測單元所提供之該檢測資訊，並依據一類神經模糊控制法則微調整該製程參數，以產生一控制命令，該控制單元傳輸該控制命令至該製程設備，該製程設備依據該控制命令執行。A sputtering process control system capable of micro-adjusting process parameters, comprising:
At least one detecting unit connected to a process device and detecting the process device to generate a detection information; and a control unit connecting the process device and the detecting unit, the control unit receiving at least one time parameter and at least one process parameter And the detection information provided by the detecting unit, and micro-adjusting the process parameter according to a type of neuro-fuzzy control law to generate a control command, the control unit transmits the control command to the process device, and the process device executes according to the control command .
匯入一時間參數、一製程參數及一檢測資訊至一濺鍍製程控制系統；
該濺鍍製程控制系統依據一類神經模糊控制法則運算該時間參數、該製程參數及該檢測資訊，以微調整該製程參數，並依據微調整之該製程參數產生一控制命令；以及
該濺鍍製程控制系統傳輸該控制命令至一製程設備，該製程設備依據該控制命令進行濺鍍製程。A sputtering process control method capable of micro-adjusting process parameters, comprising:
Importing a time parameter, a process parameter and a detection information to a sputtering process control system;
The sputtering process control system calculates the time parameter, the process parameter and the detection information according to a type of neural fuzzy control rule to finely adjust the process parameter, and generates a control command according to the micro-adjusted process parameter; and the sputtering process The control system transmits the control command to a process device, and the process device performs a sputtering process according to the control command.
該濺鍍製程控制系統檢測該製程設備進行濺鍍製程之一製程狀態，並依據該製程狀態產生更新之一檢測資訊；
該濺鍍製程控制系統依據該檢測資訊判斷該製程狀態未達到理想之製程狀態；以及
重覆執行匯入該時間參數、該製程參數及該檢測資訊至該濺鍍製程控制系統之步驟、該濺鍍製程控制系統依據該類神經模糊控制法則運算該時間參數、該製程參數及該檢測資訊，以微調整該製程參數，並依據微調整之該製程參數產生該控制命令之步驟、該製程設備依據該控制命令進行濺鍍製程之步驟及該濺鍍製程控制系統檢測該製程設備進行濺鍍製程之該製程狀態，並依據該製程狀態產生更新之該檢測資訊之步驟。The sputtering process control method as described in claim 8 of the patent application further includes:
The sputtering process control system detects a process state of the process equipment for performing a sputtering process, and generates one of the update detection information according to the process state;
The sputtering process control system determines, according to the detection information, that the process state does not reach an ideal process state; and repeatedly performs the step of importing the time parameter, the process parameter, and the detection information to the sputtering process control system, the splashing The plating process control system calculates the time parameter, the process parameter and the detection information according to the neuro-fuzzy control rule to finely adjust the process parameter, and generates the control command according to the micro-adjusted process parameter, and the process device is based on The control command performs the steps of the sputtering process and the sputtering process control system detects the process state of the process equipment for performing the sputtering process, and generates the updated detection information according to the process state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW102125164A TW201502302A (en) | 2013-07-15 | 2013-07-15 | Sputtering process control system capable of minute adjustment of process parameters and method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW102125164A TW201502302A (en) | 2013-07-15 | 2013-07-15 | Sputtering process control system capable of minute adjustment of process parameters and method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TW201502302A true TW201502302A (en) | 2015-01-16 |
| TWI495751B TWI495751B (en) | 2015-08-11 |
Family
ID=52718294
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW102125164A TW201502302A (en) | 2013-07-15 | 2013-07-15 | Sputtering process control system capable of minute adjustment of process parameters and method thereof |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TW201502302A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111534804A (en) * | 2020-06-16 | 2020-08-14 | 常州市乐萌压力容器有限公司 | Magnetron sputtering process parameter optimization method based on improved grey correlation model |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3523962B2 (en) * | 1996-05-21 | 2004-04-26 | アネルバ株式会社 | Sputtering apparatus and method for forming thin film by sputtering into hole |
| US6090246A (en) * | 1998-01-20 | 2000-07-18 | Micron Technology, Inc. | Methods and apparatus for detecting reflected neutrals in a sputtering process |
| JP4021601B2 (en) * | 1999-10-29 | 2007-12-12 | 株式会社東芝 | Sputtering apparatus and film forming method |
| WO2002063064A1 (en) * | 2001-02-07 | 2002-08-15 | Asahi Glass Company, Limited | Spatter device and spatter film forming method |
| JP3866615B2 (en) * | 2002-05-29 | 2007-01-10 | 株式会社神戸製鋼所 | Reactive sputtering method and apparatus |
| US6995545B2 (en) * | 2003-08-18 | 2006-02-07 | Mks Instruments, Inc. | Control system for a sputtering system |
| US8133359B2 (en) * | 2007-11-16 | 2012-03-13 | Advanced Energy Industries, Inc. | Methods and apparatus for sputtering deposition using direct current |
| TWI425107B (en) * | 2010-11-15 | 2014-02-01 | Ind Tech Res Inst | Continuous-type sputtering apparatus and method of fabricating solar selective absorber |
-
2013
- 2013-07-15 TW TW102125164A patent/TW201502302A/en unknown
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111534804A (en) * | 2020-06-16 | 2020-08-14 | 常州市乐萌压力容器有限公司 | Magnetron sputtering process parameter optimization method based on improved grey correlation model |
Also Published As
| Publication number | Publication date |
|---|---|
| TWI495751B (en) | 2015-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112884136B (en) | Bounded clustering projection synchronous regulation control method and system for coupled neural network | |
| Li et al. | Non-fragile state estimation for delayed fractional-order memristive neural networks | |
| Mia et al. | An algorithm for training multilayer perceptron (MLP) for Image reconstruction using neural network without overfitting | |
| CN109100935B (en) | Intelligent damping PI control method of large-time-lag system | |
| Tian et al. | Time-delay compensation method for networked control system based on time-delay prediction and implicit PIGPC | |
| Jau et al. | Modified quantum-behaved particle swarm optimization for parameters estimation of generalized nonlinear multi-regressions model based on Choquet integral with outliers | |
| CN108460462A (en) | A kind of Interval neural networks learning method based on interval parameter optimization | |
| CN114567288B (en) | Distribution collaborative nonlinear system state estimation method based on variable decibels | |
| CN111832693A (en) | Neural network layer operation and model training method, device and equipment | |
| TW201502302A (en) | Sputtering process control system capable of minute adjustment of process parameters and method thereof | |
| CN112180733B (en) | Fuzzy logic-based building energy consumption system prediction control parameter setting method | |
| CN107957685B (en) | Neurodynamics method for solving noise-containing time-varying problem | |
| Li et al. | Predicting Software Quality by Optimized BP Network Based on PSO. | |
| Du et al. | A novel locally regularized automatic construction method for RBF neural models | |
| JP2002042107A (en) | Learning method for neural network | |
| CN106773672A (en) | Improve the new three ranks linear extended state observer building method of accuracy of observation | |
| CN105527841A (en) | Networking tracking control method of time-varying signal | |
| CN110598226A (en) | Nonlinear system construction method based on collective estimation and neural network | |
| Sun et al. | Nonlinear function approximation based on least Wilcoxon Takagi-Sugeno fuzzy model | |
| Li et al. | Parameter estimation of multiple‐input single‐output Hammerstein controlled autoregressive system based on improved adaptive moment estimation algorithm | |
| Jia et al. | The probability density function based neuro-fuzzy model and its application in batch processes | |
| CN115870350A (en) | Prediction method of plate strip convexity, intelligent roller control method and computing equipment | |
| Xu et al. | Support Vector Regression Model Predictive Control Based on LM Algorithm and BA | |
| Shen et al. | Stock price prediction based on optimized BP neural network | |
| Kusumoputro et al. | An improved single neuron adaptive PID controller system based on additional error of an inversed-control signal |