Real Time Etch-depth Measurement Using Surface Acoustic Wave Sensor

Transcription

1 Australian Journal of Basic and Applied Sciences, (8): -7, 1 ISSN Real Tie Etch-depth Measureent Using Surface Acoustic Wave Sensor 1 Reza Hosseini, Navid Rahany, 3 Behrad Soltanbeigi, Rouzbeh Aghabeighi Harris 1,3, Departent of Electrical Engineering, Khoy Branch, Islaic Azad University, Khoy, Iran Departent of Electrical Engineering, Tafresh Branch, Islaic Azad University, Tafresh, Iran Abstract: Measuring the depth of etching layers is one of the ajor difficulties in fabrication process, considering the iproveents ade in electronically devices anufacturing, which coonly takes place after etching stage. In this paper, it has been tried to do the easureent of etching layer depth siultaneously and parallel with the etching stage, using Surface Acoustic Wave (SAW) sensors which have the ability to easure all kinds of physical paraeters. The way of this sensor perforance and its design, based on different available paraeters and the study of this design accuracy, are the ain purposes of this paper. Key words: piezoelectric, interdigital transducers, SAW sensor, etching INTRODUCTION For aking integrated circuits with iniu features size, we ll need exact control of different anufacturing paraeters. There are various ethods for controlling and easuring this paraeters but ost of the don t take place siultaneously. This eans that after etching stage, depth of etched area is easured through particular devices or systes. So, we can t have exact control over this stage. The ethods which can show the depth of etching area together with iproveent of the stage and so we can have exact and ore efficient control over this stage, are of iportance. We can use different sensors for ipleentation of these ethods. One of these sensors is SAW. These sensors active through surface acoustic waves and can easure the variations by varying aplitude or waves phases( Draft.B, 1; Morgan David P, 1998; Yudistira et al., 9; heribsek et al., 1). How we can use these waves and design a sensor to be used for controlling and displaying the depth of etching area, is studied in this paper and a sensor is designed for this purpose. SensorPperforance: SAW sensor used in the design is an oscillator delay line stabilized (Crrab et al, 1973) which has covered diffusion path with thin layer of aterial which is due to be etched. During the etching by taking this aterial, the ass covered delay line decreases and causes the increase of oscillator frequency (Lee et al., ). So by having an oscillator with reference frequency and coparing it with the above entioned oscillator frequency, we can copute reduced ass. This eans that the sensor is coposed of two SAW oscillator devices. In this paper, we have used photoresist for covering sensor that the thickness of this layer is ultiately easured while etching. 3 Sensor Design: For designing SAW sensor, we should consider its two ain parts: 1) piezoelectric substrate; ) interdigital transducers (IDT) ( Wang et al. ), that choosing aterials used for both parts and designing IDT is studied in this part (Slobodink et al., 197; Ballantine et al., 1997). 3.1 Choosing Piezoelectric Substrate: The ost suitable piezoelectric aterials which are used in SAW sensor are: Y-Z, Y-X,, ST-Quartz, Y-Z, aong which Quartz and Lithiu Niobate are of ost use in aking SAW devices as there are rather thorough inforation about their features. For designing the expected SAW sensor, piezoelectric substrate Lithiu Niobate has particular significance despite its high iportance because of high coupling coefficient, as it has less signal distortion and losses and the accuracy of designed sensor rises regarding SAW device application as a sensor. 3. Choosing suitable aterial for IDT deposit: For IDT deposit in SAW devices, Aluinu and gold are often used. We choose Aluinu as it sticks the surface of substrate well, has fair price, is deposited easily, has good condition, and has less density. Corresponding Author: Reza Hosseini, Departent of Electrical Engineering, Khoy Branch, Islaic Azad University, Khoy, Iran E-ail: hosseini@iaukhoy.ac.ir

2 Aust. J. Basic & Appl. Sci., (8): -7, Design Coputations: Considering available lithographic facilities, and as 5µ line width can be ade well, we have the following for the expected devices: Wf = λ = 5µ λ = d = 5µ = 1µ So, the width of each finger for producing expected central frequency and center space to digital coupled finger center fro two IDT electrode is. Central frequency in substrate can be coputed as: ν f = λ where v=388 /sec is Wave velocity in substrate and λ is the wavelength. So, we have: ν 388 f = = = 3.88MHz λ 1µ Nuber of fingers per electrode and bandwidth of SAW sensor can be coputed as: π π N = = = 3.75 N = 5.7 N P k.8 P P f 3.88MHz BW. = = = 5.81MHz NP where k =.8% and is electroechanical coupling coefficient. For getting the best reply fro the expected device, interdigital transducer ipedance should atch with easureent systes ipedance. After designing if input and output ipedance of the device don t atch with the systes connected to it, we should use external atching circuits for collating together. If we choose IDT ipedance for atching with easuring devices and transferring utost power in 5 ῼ central ipedance: 1 K N W =. P Z f c s π + ( K NP ) W = W 11 where C s = F/. as doesn't have any effects on the design, we suppose it six fold each finger width: Wch = 5µ = 15µ h = W + Wch = The other paraeter is center-to-center space of two transducers in SAW sensor. It is chosen in away that scattering losses and direct electroagnetic coupling between two IDTs becoe the least. For having less scattering losses and also less ebedded losses, two IDTs should be close to each other. But this causes the rise of probable direct electroagnetic coupling between two transducers. The aount entioned for this paraeter is usually one hundred fold to three-hundred fold of wave length. We consider one hundred fold of wave length in this part. In the table 1, the paraeters related to the design have coe altogether. According to these paraeters, the susceptance, inductance and ipedance of SAW sensor on Y-Z LiNbO3 Substrate are shown in figure 1. Table 1: paraeters used to IDT on Y-Z LiNbO3 Substrate. IDT Paraeters Nuber of fingers per electrode (N p ) Fingers width (W f ) Spacing width (W s ) Fingers period length (d) Distance between IDT centers(l) Acoustic aperture(w) IDT total length(h) IDT channel width (W ch ) Value 5μ 5μ 1μ μ 7

3 Aust. J. Basic & Appl. Sci., (8): -7, 1 Ga(w) or Ba(w) Or Yt(w) For Y-Z LiNbO3 1 x Ga(w) - Ba(w) Yt(w) x Y-Z LiNbO3 Yt(w) For Y-Z LiNbO Fig. 1: The susceptance, inductance and ipedance of SAW sensor on Y-Z LiNbO3 Substrate. Also, we can do the design for Lithiu Niobate, i.e., Y-X. Design values are in the table. According to these values, we can obtain the results which shown in Figure. Experients: Because of ass loading, the frequency of covered oscillator are less than the frequency of reference oscillator [9]. By using approxiate equation, we can tell: ( f1 f ) = f = ( k1 + k ) hρf1 that and are aterial constant, is density, h is photoresist layer. In this experient, h= and ρ=1 3 kg/ 3. Also we know that for Lithiu Niobate (k 1 +k )= sec/kg. For Y-Z : f1 = 3.88MHz f = KHz f = 3.718MHz A shift of electrical phase is added to the ring oscillator which is covered with photoresist so that it decreases, as uch as 15 KHz. The purpose of exerting this phase shift is certainty of the fact that even after photoresist etching, doesn t becoe zero: f = 3.718MHz 15KHz = 3. 58MHz So at the beginning of etching stage and easuring sensor. In the sae way for 18 Y-X LiNbO3: f = 39.9MHz f = 1.19KHz f = 39.78MHz f = 39.78MHz 15KHz = MHz 8

4 Aust. J. Basic & Appl. Sci., (8): -7, 1 Table : paraeters used to IDT on 18 Y-X LiNbO 3 Substrate IDT Paraeters Nuber of fingers per electrod (N p ) Fingers width (W f ) Spacing width (W s ) Fingers period length (d) Distance between IDT centers(l) Acoustic aperture(w) IDT total length(h) IDT channel width (W ch ) Value 5 5μ 5μ 1μ μ Ga(w) or Ba(w) Or Yt(w) For 18 Y-X LiNbO3 1 x Ga(w) - Ba(w) Yt(w) Fig. : The susceptance, inductance and ipedance of SAW sensor on 18 Y-X LiNbO 3 Substrate. Therefore, at the beginning of etching stage and easuring sensor. Frequencies of reference oscillator and oscillator covered with photoresist, follow each other by teperature, so is alost independent of teperature and varies only with photoresist layer thichness. Frequency of reference oscillator gives us teperature rate of substrate. has been drawn in ters of photoresist layer thickness in figure 3 during etching process. As we see in figure 3, after coplete etching and photoresist reoval, reains constant over 15KHz. versus etching tie, has been shown in figure. We notice that variations in τ 1 < t < τ,that τ 1 =15in, τ =in, is linear. x 15 Y-ZLiNbO3 18Y-XLiNbO3 3.5 Defference Frequency The Thickness of Photoresist fil x 1-7 Fig. 3: deference of frequency versus the Thickness of photoresist fil during etching process For < t < τ. the speed of variations is uch saller than τ 1 < t < τ.this can be explained by figure 5 that shows the typical variations of SAW sensor speed as a function of for average layers consisting of a slow layer (photoresist) over a fast layer ( ). 9

5 Aust. J. Basic & Appl. Sci., (8): -7, Deference of frequency (khz) Tie(in) Fig. : versus etching tie Here, h is photoresist layer thickness, and λ is SAW wave length. We see that for sall aount speed variations are linear. By increase, graph slope decreases. In our etching experient, in t= has the ost aount and decreases by tie increase. So, at the beginning a certain variations in fil thickness will ake a sall variation in SAW speed. Considering, variation in speed akes soe variation in oscillator frequency. Although etching rate can be steady, chart of tie will have saller slope for saller aount of tie. At the end of in, alost the whole photoresist layer has been reoved. The survey of device in 5 in shows pieces of resistant that are soe parts of diffusion site. For, photoresist layer has been copletely reoved, as a result alost reains constant (Willias et al., 1997). Conclusion: The sallest variation that can take place in photoresist thickness has been liited by short tie stability of SAW oscillator. Short tie stabilities better than, can be achieved by SAW delay lines oscillator and is also possible by SAW oscillator instead of delay line. For having a glipse, we ust to study short-ter stability of. This eans that variations can be. In the present device, decreases in fil thickness, creates a frequency shift of 93KHz. This is in accordance with variation in for variation in thickness. A short tie stable oscillator can yield.1n accuracy. REFERENCES Ballantine et al., Acoustic wave Sensors. Theory, Design, And Physico-Cheical Application. Acadeic Press. Crrab et al., Surface Acoustic Wave Oscillators : Mode Selection And Frequency Modulation. Electronics Lett, 9(1): Drafts, B., 1. Acoustic Wave Technology Sensors. IEEE Transactions on Microwave Theory and Techniques, 9( ): heribsek et al., 1, Surface Acoustic Wave Sensors in Mechanical Engineering. FME Transaction, 38(1): Lee et al.,. Real tie onitoring of thickness of silicon ebrane during wet etching using a novel surface acoustic wave sensor. Proceedings of the SPIE, pp: Morgan David, P., History of SAW Device. Proceeding of The IEEE Internatioal Frequency Control Syposiu., pp: 39-. Slobodink, Andrew j, 197. Surface Acoustic Wave And SAW Materials. Proceeding of the IEEE, (5): Wang et al.. Field analysis and calculation of interdigital transducers with arbitrary finger shapes. J. Phys. D: Appl. Phys. 39(): 9. Willias, K. and R. Muller, 199. Etch rates for icroachining processing. Journal of Microelectroechanical systes, 5(): 5-9. Yudistira et al., 9. Surface acoustic wave generation in ZX-cut LiNbO 3 superlattices using coplanar electrodes. Appl. Phys. Lett., 95: