TWI731851B - Method and device for cleaning semiconductor substrate - Google Patents

Abstract

A method for cleaning a semiconductor substrate without damaging the patterned structure on the semiconductor substrate using a super/megasonic device includes spraying liquid into the gap between the semiconductor substrate and the super/megasonic device; setting a super/megasonic power supply The frequency is f1 and the power is P1 to drive the ultra/megasonic device; before the cavity oscillation in the liquid damages the patterned structure on the semiconductor substrate, the frequency of the ultra/megasonic power supply is set to f ₂ and the power is P ₂ To drive the ultra/megasonic device; after the temperature in the bubble cools to the set temperature, set the frequency of the ultra/megasonic power supply to f ₁ and the power to P _{1 again} ; repeat the above steps until the semiconductor substrate is cleaned. Generally, if f ₁ = f ₂ , then P ₂ is equal to 0 or much less than P ₁ ; if P ₁ = P ₂ , then f _{2 is} greater than f ₁ ; if f ₁ <f ₂ , then P ₂ can be equal to or less than P ₁ .

Description

Method and device for cleaning semiconductor substrate

The present invention relates to a method and device for cleaning a semiconductor substrate, in particular to controlling the cavitation oscillation generated by an ultrasonic/megasonic device during the cleaning process to obtain stable or controllable cavitation oscillation on the entire substrate, effectively removing particles, Without damaging the device structure on the substrate.

Semiconductor devices are formed by forming transistors and interconnections on a semiconductor substrate through a series of different processing steps. Recently, the establishment of transistors has changed from two-dimensional to three-dimensional, such as fin-type field-effect transistors. In order to electrically connect the transistor terminal and the semiconductor substrate, conductive (for example, metal) grooves, holes, and other similar structures must be made on the dielectric material of the semiconductor substrate as a part of the device. Slots and holes can transmit electrical signals and energy between transistors, internal circuits, and external circuits.

In order to form a fin-type field effect transistor and an interconnection structure on a semiconductor substrate, the semiconductor substrate needs to go through multiple steps, such as masking, etching, and deposition, to form the required electronic circuits. In particular, the multi-layer mask and plasma etching steps can form fin-type field effect transistors and/or patterns of recessed regions in the dielectric layer of the semiconductor substrate as the fins of the transistors and/or the grooves and through holes of the interconnection structure. . In order to remove the fin structure and the ash during the etching or photoresist ashing process / Or particles and contamination generated in the grooves and through holes must be wet cleaned. In particular, when the device manufacturing node is constantly approaching or smaller than 14 or 16 nm, the sidewall loss of fins and/or grooves and vias is the key to maintaining critical dimensions. In order to reduce or eliminate sidewall loss, it is important to use mild, diluted chemical reagents, or sometimes only deionized water. However, diluted chemical reagents or deionized water generally cannot effectively remove particles in the fin structure and/or grooves and through holes. Therefore, it is necessary to use mechanical force to effectively remove these particles, such as ultrasonic waves/megasonic waves. Ultrasonic/megasonic waves will generate cavitation oscillations to provide mechanical force to the substrate structure. These violent cavitation oscillations such as unstable cavitation oscillations or micro-jets will damage these patterned structures. Maintaining stable or controllable cavitation oscillation is a key parameter to control the limit of mechanical damage and effectively remove particles.

It is mentioned in U.S. Patent No. 4,326,553 that megasonic energy can be used in combination with nozzles to clean semiconductor substrates. The fluid is pressurized, and megasonic energy is applied to the fluid through the megasonic sensor. A nozzle of a specific shape ejects a liquid like a ribbon, which vibrates at a megasonic frequency on the surface of the substrate.

In U.S. Patent No. 6,039,059, it is mentioned that an energy source transmits sound wave energy into the fluid by vibrating an elongated probe. In one example, the fluid is sprayed on the front and back sides of the substrate, and a probe is placed close to the upper surface of the substrate. In another example, the end of a short probe is placed close to the surface of the substrate. During the rotation of the substrate, the probe moves on the surface of the substrate.

In U.S. Patent No. 6,843,257 B2, it is mentioned that an energy source causes a rod to vibrate about an axis parallel to the surface of the substrate. The surface of the rod is etched into curvilinear dendritic shapes, such as spiral grooves.

In order to effectively remove particles without damaging the device structure on the substrate, a good method is needed to control the cavitation oscillation generated by the ultrasonic/megasonic device during the cleaning process to obtain stable or controllable air on the entire substrate. Cavitation.

The present invention proposes a method to achieve no damage to the patterned structure on the substrate by maintaining stable air cavity oscillation when cleaning the substrate using ultrasonic/megasonic waves. Stable cavitation oscillation can be achieved by setting the power of the sonic power supply to P _{1 when the time interval is less than τ 1} , and setting the power of the sonic power supply to P _{2 when the time interval is greater than τ 2.} Repeat the above steps until the substrate is cleaned, where the power P ₂ is equal to 0 or much less than the power P ₁ , τ ₁ is the time interval for the temperature in the bubble to rise to the critical implosion temperature, and τ ₂ is the time interval for the temperature in the bubble to drop far below the critical implosion temperature.

The present invention proposes another method of using ultrasonic/megasonic waves to clean the substrate by maintaining stable air cavity oscillation to achieve no damage to the patterned structure on the substrate. Stable cavitation oscillation can be achieved by setting the frequency of the _{sonic power supply to f 1 in the time interval less than τ 1} and setting the frequency of the sonic power supply to f ₂ in the time interval greater than τ _2. Repeat the above steps until the substrate is cleaned, where f ₂ Much greater than f ₁ , preferably 2 or 4 times _{of f 1 , τ 1} is the time interval for the temperature in the bubble to rise to the critical implosion temperature, and τ ₂ is the temperature in the bubble falling far below the critical implosion temperature Time interval.

The present invention also proposes a method to achieve no damage to the patterned structure on the substrate by maintaining stable air cavity oscillation when cleaning the substrate using ultrasonic/megasonic waves, and the size of the bubbles is smaller than the spacing between the patterned structures. Stable bubble having a size smaller than the spacing between the patterned structure is provided through acoustic cavitation oscillation power is less than a time interval τ ₁ ratio internal strength P _1, is provided at the acoustic power is greater than the time interval τ ₂ internal strength ratio P _2, repeat the above steps Until the substrate is cleaned, where the power P ₂ is equal to 0 or much less than the power P ₁ , τ ₁ is the time interval for the bubble size to increase to a critical size, which is equal to or greater than the spacing between the patterned structures, τ ₂ is the time interval for the size of the bubbles to decrease to a value much smaller than the spacing between the patterned structures.

The present invention also proposes a method to achieve no damage to the patterned structure on the substrate by maintaining stable air cavity oscillation when cleaning the substrate using ultrasonic/megasonic waves, and the size of the bubbles is smaller than the spacing between the patterned structures. Stable bubble having a size smaller than the spacing between the patterned structure is provided through acoustic cavitation oscillation power during a time interval τ ₁ is less than the frequency f _1, is provided at the acoustic power is greater than the time interval [tau] ₂ of frequency f _2, repeat the above steps Until the substrate is cleaned, where f _{2 is} much larger than f ₁ , preferably 2 or 4 times _{of f 1 , τ 1} is the time interval for the bubble size to increase to the critical size, which is equal to or larger than the pattern The distance between the patterned structures, τ ₂ is the time interval at which the size of the bubble decreases to a value much smaller than the distance between the patterned structures.

1003‧‧‧Ultrasonic/Megasonic Device

1004‧‧‧Piezoelectric sensor

1008‧‧‧Acoustic Resonator

1010‧‧‧wafer

1012‧‧‧Nozzle

1014‧‧‧wafer chuck

1016‧‧‧Drive device

1032‧‧‧Deionized water (cleaning liquid chemical reagent)

3003‧‧‧Ultrasonic/Megasonic Device

4034‧‧‧Fine structure

6080‧‧‧Micro nozzle

6082‧‧‧Bubble

15010‧‧‧wafer

15034‧‧‧Pattern structure

15046‧‧‧Bubble

15048‧‧‧Bubble

16010‧‧‧wafer

16014‧‧‧wafer chuck

16016‧‧‧Drive device

16060‧‧‧Deionized water (cleaning liquid chemical reagent)

16062‧‧‧Ultrasonic/Megasonic Device

16064‧‧‧Nozzle

17010‧‧‧wafer

17072‧‧‧Ultrasonic/Megasonic Device

17070‧‧‧Cleaning liquid chemical reagent

17074‧‧‧Solution tank

17076‧‧‧wafer box

1A-1B are exemplary embodiments of a wafer cleaning device using an ultrasonic/megasonic device; FIGS. 2A-2G are various shapes of ultrasonic/megasonic sensors; Figure 3 shows the cavitation oscillation during the wafer cleaning process; Figures 4A-4B show the patterned structure on the wafer damaged by the unstable cavitation oscillation during the cleaning process; Figure 5A-5C shows the thermal energy inside the bubble during the cleaning process Changes; FIGS. 6A-6C are an exemplary embodiment of a wafer cleaning method; FIGS. 7A-7C are another exemplary embodiment of a wafer cleaning method; FIGS. 8A-8D are another exemplary embodiment of a wafer cleaning method 9A-9D is another exemplary embodiment of a wafer cleaning method; FIGS. 10A-10B are another exemplary embodiment of a wafer cleaning method; FIGS. 11A-11B are another exemplary implementation of a wafer cleaning method Examples; FIGS. 12A-12B are another exemplary embodiment of a wafer cleaning method; FIGS. 13A-13B are another exemplary embodiment of a wafer cleaning method; FIGS. 14A-14B are another exemplary embodiment of a wafer cleaning method 15A-15C are stable cavitation oscillation damage patterned structure on the wafer during the cleaning process; FIG. 16 is another exemplary embodiment of a wafer cleaning device using an ultrasonic/megasonic device; FIG. 17 Figures 18A-18C are another exemplary embodiment of a wafer cleaning method; Figure 19 is another exemplary embodiment of a wafer cleaning method.

In order to make the above objectives, features and advantages of the present invention more capable Obviously and understandably, the specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings to make the above and other objectives, features and advantages of the present invention clearer. The drawings are not drawn to scale deliberately, and the focus is on showing the gist of the present invention.

Figures 1A-1B illustrate a wafer cleaning device using an ultrasonic/megasonic device. The wafer cleaning device includes a wafer 1010, a wafer chuck 1014 driven and rotated by a rotating drive device 1016, a nozzle 1012 spraying cleaning liquid chemicals or deionized water 1032, an ultrasonic/megasonic device 1003, and an ultrasonic/megasonic power supply . The ultrasonic/megasonic device 1003 further includes a piezoelectric sensor 1004 and an acoustic resonator 1008 paired therewith. The sensor 1004 vibrates after being energized, and the resonator 1008 transfers high-frequency sound energy into the liquid. The cavitation oscillation generated by the ultrasonic/megasonic energy loosens the particles on the surface of the wafer 1010, so that the contaminants are separated from the surface of the wafer 1010, and then are removed from the surface of the wafer through the flowing liquid 1032 provided by the shower head 1012.

Figures 2A-2G illustrate top views of the ultrasonic/megasonic device of the present invention. The ultrasonic/megasonic device 1003 shown in FIG. 1 can be replaced by an ultrasonic/megasonic device 3003 of different shapes, such as a triangle or pie shape as shown in FIG. 2A, a rectangle as shown in FIG. 2B, and an eighth as shown in FIG. 2C. The polygonal shape is the ellipse shown in Fig. 2D, the semicircle shown in Fig. 2E, the quarter circle shown in Fig. 2F, and the circle shown in Fig. 2G.

Figure 3 illustrates cavitation oscillations during compression. The shape of the bubble is gradually compressed from spherical A to apple-shaped G, and finally the bubble reaches the implosion state I and forms a microjet. As shown in FIGS. 4A and 4B, micro-jetting is very violent (can reach thousands of atmospheres and thousands of degrees Celsius), which can damage the fine structure 4034 on the semiconductor wafer 1010, especially when the feature size is reduced to 70 nm and less.

Figures 5A-5C illustrate simplified models of cavitation oscillations of the present invention. When the positive pressure of the sound wave acts on the bubble, the bubble reduces its volume. In the volume reduction process, the acoustic pressure P _M is the bubble work, mechanical work converted to heat inside of the bubble, and therefore, the gas inside the bubble and / or the temperature of the steam increases.

The ideal gas equation can be expressed as follows: p ₀ v ₀ /T ₀ =pv/T (1)

Among them, P ₀ is the pressure inside the bubble before compression, V ₀ is the initial volume of the bubble before compression, T ₀ is the gas temperature inside the bubble before compression, P is the pressure inside the bubble under pressure, and V is the pressure inside the bubble under pressure. Volume, T is the gas temperature inside the bubble under pressure.

In order to simplify the calculation, it is assumed that the temperature of the gas does not change when the compression or compression is very slow. Since the liquid surrounds the bubbles, the temperature increase can be ignored. Thus, a bubble during the compression (volume from 1 to N unit amount per unit volume or a compression ratio of the amount of N), the sound pressure P _M mechanical work done W _m can be expressed as ^{_{follows: w m = ʃ 0 x0-1 pSdx}} = ʃ ₀ ^x0-1 (S(x ₀ p ₀ )/(x ₀ -x))dx=Sx ₀ p ₀ ʃ ₀ ^x0-1 dx/(x ₀ -x)=-Sx ₀ p ₀ ln(x ₀ -x)| ₀ ^x0-1 =Sx ₀ p ₀ ln(x ₀ ) (2)

Among them, S is the area of the cylinder section, x ₀ is the length of the cylinder, and p ₀ is the pressure of the gas in the cylinder before compression. Equation (2) does not consider the temperature increase in the compression process. Therefore, due to the increase in temperature, the actual pressure in the bubble will be higher. In fact, the mechanical work done by the sound pressure is greater than the value calculated by the equation (2).

Suppose that the mechanical work done by sound pressure is partly converted into heat energy, and partly converted into mechanical energy of high-pressure gas and steam in the bubble. These heat energy completely promotes the increase of the gas temperature inside the bubble (no energy is transferred to the liquid molecules around the bubble), assuming that before and after compression The quality of the gas in the bubble remains unchanged. After the bubble is compressed once, the temperature increase T can be expressed by the following equation: T=Q/(mc)=β w _m /(mc)=β Sx ₀ p ₀ ln(x ₀ ) /(mc) (3) Among them, Q is the thermal energy converted from mechanical work, β is the ratio of the total mechanical work done by thermal energy and sound pressure, m is the gas mass in the bubble, and c is the specific heat coefficient of the gas.

Set β=0.65, S=1E-12m ² , x ₀ =1000 m=1E-3m (compression ratio N=1000), p ₀ =1kg/cm ² =1E4kg/m ² , m=8.9E-17kg( For hydrogen), c=9.9E3 J/(kg ⁰ k) is substituted into equation (3), then ΔT=50.9 ⁰ k.

_{The gas temperature T 1} in the bubble after one compression can be calculated: T ₁ =T ₀ +ΔT=20℃+50.9℃=70.9℃ (4)

When the bubble reaches the minimum value of 1 micron, as shown in Figure 5B. At such a high temperature, the liquid around the bubble evaporates, and then the sound pressure becomes negative, and the bubble starts to increase. In this reverse process, _{the hot gas and steam with pressure P G} will do work on the surrounding liquid surface. Meanwhile, the sound pressure P _M stretching bubble toward the direction of expansion, shown in Figure 5C. Therefore, the negative sound pressure _PM also does some work on the surrounding liquid. As a result of the interaction, the heat energy in the bubble cannot be completely released or converted into mechanical energy. Therefore, the gas temperature in the bubble cannot be reduced to the initial gas temperature T ₀ or the liquid temperature. As shown in FIG. 6B, after the first period of cavitation oscillation is completed, the gas temperature T ₂ in the bubble will be between T ₀ and T ₁ . T ₂ can be expressed as follows: T ₂ =T1-δT=T ₀ +ΔT-δT (5)

Among them, δT is the temperature decrease after the bubble expands once, and δT is less than ΔT.

When the second period of cavitation oscillation reaches the minimum bubble size, the temperature T3 of the gas or vapor in the bubble is: T3=T2+ΔT=T ₀ +ΔT-δT+ΔT=T ₀ +2ΔT-δT (6)

When the second period of cavitation oscillation is completed, the temperature T4 of the gas or vapor in the bubble is: T4=T3-δT=T ₀ +2ΔT-δT-δT=T ₀ +2ΔT-2δT (7)

Similarly, when the n-th period of cavitation oscillation reaches the minimum bubble size, the temperature T _2n-1 of the gas or steam in the bubble is: T _2n-1 =T ₀ +nΔT-(n-1)δT (8 )

When the nth cycle of cavitation oscillation is completed, the temperature T _2n of the gas or steam in the bubble is: T _2n = T ₀ +nΔT-nδT=T ₀ +n(ΔT-δT) (9)

As the number n of cavitation oscillation cycles increases, the temperature of the gas and steam will also increase. Therefore, the more molecules on the bubble surface evaporate into the bubble 6082, the bubble 6082 will also become larger, as shown in FIG. 6C. Finally, the temperature of the compressed gas bubbles within the implosion process will reach a temperature T _i (T _i temperature implosion typically up to several thousand degrees Celsius) to form a heavy microprojection 6080, shown in Figure 6C.

According to formula (8), the number of implosion cycles n _i can be expressed as follows: n _i =(T _i =T ₀ -ΔT)/(ΔT-δT)+1 (10)

According to the formula (1()), the implosion time t ₁ can be expressed as follows: τ _i =n _i t ₁ =t ₁ ((T _i =T ₀ -ΔT)/(ΔT-δT)+1)=n _i / f ₁ =((T _i =T ₀ -ΔT)/(ΔT-δT)+1)/f ₁ (11)

Among them, t ₁ is the cycle period, and f ₁ is the frequency of ultrasound/megasonic waves.

According to formulas (10) and (11), the number of implosion cycles n _i and the implosion time 壜_i can be calculated. Table 1 shows the relationship between the number of implosion cycles n _i , the implosion time t ₁ and (Δ T- δ T), assuming Ti=3000℃, Δ T=50.9℃, T ₀ =20℃, f ₁ =500KHz, f ₁ = 1MHz, and f ₁ = 2MHz.

In order to avoid damage to the patterned structure on the wafer, it is necessary to maintain stable cavitation oscillation to avoid micro-jetting caused by bubble implosion. FIGS. 7A-7C show the patterned structure on the wafer without damaging the patterned structure by maintaining stable air cavity oscillation when cleaning the wafer with ultrasonic/megasonic waves proposed by the present invention. Fig. 7A is the output waveform of the power supply; Fig. 7B is the temperature curve corresponding to each cavitation oscillation period; Fig. 7C is the expansion size of the bubble corresponding to each cavitation oscillation period. The operation process steps for avoiding bubble implosion according to the present invention are as follows:

Step 1: Place the ultrasonic/megasonic device near the surface of the wafer or substrate set on the chuck or solution tank;

Step 2: Fill the space between the wafer and the ultrasonic/megasonic device with chemical liquid or gas with water (hydrogen, nitrogen, oxygen or carbon dioxide);

Step 3: Spin the chuck or vibrate the wafer;

Step 4: Set the power supply frequency to f ₁ and power to P ₁ ;

Step 5: Before the gas or steam temperature in the bubble reaches the implosion temperature T _i (or the time reaches τ ₁ <τ _i , τ _i is calculated by formula (11)), set the output power of the power supply to 0 watts, therefore, Since the temperature of the liquid or water is much lower than the temperature of the gas, the temperature of the gas inside the bubble begins to drop.

Step 6: After the gas temperature in the bubble is reduced to normal temperature T ₀ or the time (time of zero power) reaches τ ₂ , set the power frequency to f ₁ and the power to P _{1 again} .

Step 7: Repeat step 1 to step 6 until the wafer is cleaned.

In step 5, in order to avoid the bubble burst, it must be less than the time τ ₁ τ _i, τ _i can be calculated by the formula (II). In step 6, the temperature of the gas in the bubble does not have to be cooled to the normal temperature or the temperature of the liquid. It can be a specific temperature higher than the normal temperature or the temperature of the liquid, but it is better to be much lower than the implosion temperature _i .

According to formulas 8 and 9, if (Δ T- δ T) is known, τ _i can be calculated. But generally speaking, (Δ T- δ T) is not easy to be calculated or obtained directly. The implosion time τ _i can be obtained through experiments in the following steps.

Step 1: Based on Table 1, select five different times τ ₁ as the conditions set by the DOE experiment;

Step 2: Choose a time τ _{2 that is at least ten times τ 1} , preferably 100 times τ ₁ in the first test.

Step 3: Use the determined power P _{0 to} run the above five conditions to clean the wafers with the patterned structure respectively. Here, P ₀ is the power determined to cause damage to the patterned structure of the wafer in a continuous uninterrupted mode (non-pulse mode).

Step 4: Use SEMS or wafer pattern damage viewing tools to check the damage of the above five types of wafers, such as AMAT SEM view or Hitachi IS3000, and then the implosion time τ _i can be determined within a certain range.

Repeat steps 1 to 4 to narrow the range of implosion time τ _i . Knowing the implosion time τ _i , τ _i can be set to a value _{less than 0.5τ i under the safety factor.} The following is an example to describe the experimental data: the patterned structure is a 55nm polysilicon grid line, the ultrasonic/megasonic frequency is 1MHZ, and the ultrasonic/megasonic device manufactured by Prosys is used in a pitch oscillation mode (PCT/CN2008/073471 publication) Operate to achieve better uniform energy within and between wafers. Table 2 below summarizes other test parameters and the final pattern damage data:

It can be seen from the above table that under the feature size of 55nm, when τ ₁ =2ms (or the number of cycles is 2000), the damage to the patterned structure is as high as 1216 points; but τ ₁ =0.1ms (or the number of cycles is At 100), the damage to the patterned structure is zero. Therefore, τ ₁ is a value between 0.1 ms and 2 ms, and further experiments are needed to narrow this range. Obviously, the number of cycles is related to the power density and frequency of ultrasonic/megasonic waves. The greater the power density, the smaller the number of cycles; the lower the frequency, the smaller the number of cycles. From the above experimental results, it can be predicted that the number of cycles without damage should be less than 2000, assuming that the power density of ultrasonic/megasonic waves is greater than 0.1w/cm ² and the frequency is less than or equal to 1MHZ. If the frequency is increased to greater than 1MHZ or the power density is less than 0.1w/cm ² , it can be predicted that the number of cycles will increase.

Knowing the time τ ₁ , τ ₂ can be shortened based on the DEO method similar to the above. Determine the time τ ₁ and gradually shorten the time τ ₂ to run the DOE until it can be observed that the patterned structure is damaged. As the time τ ₂ is shortened, the temperature of the gas or steam in the bubble cannot be sufficiently cooled, which will cause the average temperature of the gas or steam in the bubble to rise gradually, and eventually trigger the bubble implosion. The trigger time is called critical cooling time. After knowing the critical cooling time τ _c , in order to increase the safety factor, the time τ ₂ can be set to a value _{greater than 2τ c.}

Figures 8A-8D illustrate a method for cleaning a wafer using an ultrasonic/megasonic device according to the present invention. This method is similar to the method illustrated in FIG. 7A, except that in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₁ , and the power is a waveform whose amplitude changes. FIG. 8A illustrates another cleaning method that in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₁ , and the power has a waveform with increasing amplitude. FIG. 8B illustrates that another cleaning method is to set the frequency of the ultrasonic/megasonic power source to f ₁ in step 4, and the power has a waveform with a decreasing amplitude. FIG. 8C illustrates that another cleaning method is to set the frequency of the ultrasonic/megasonic power supply to f ₁ in step 4, and the power has a waveform whose amplitude first decreases and then increases. FIG. 8D illustrates that another cleaning method is to set the frequency of the ultrasonic/megasonic power supply to f ₁ in step 4, and the power has a waveform whose amplitude first increases and then decreases.

Figures 9A-9D illustrate a method of cleaning a wafer using an ultrasonic/megasonic device according to the present invention. This method is similar to the method illustrated in FIG. 7A, except that step 4 sets the frequency of the ultrasonic/megasonic power supply to a constantly changing frequency. FIG. 9A illustrates another cleaning method that in step 4, the frequency of the ultrasonic/megasonic power supply is set to f _{1 first} , and then f ₃ , and f _{1 is} higher than f ₃ . FIG. 9B illustrates another cleaning method that in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₃ first, then f ₁ , and f _{1 is} higher than f ₃ . Fig. 9C illustrates another cleaning method that in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₃ first, then f ₁ , and finally f ₃ , and f _{1 is} higher than f ₃ . Fig. 9D illustrates another cleaning method that in step 4, the frequency of the ultrasonic/megasonic power supply is set to f _{1 first} , then f ₃ , and finally f ₁ , and f _{1 is} higher than f ₃ .

Similar to the method illustrated in FIG. 9C, in step 4, the frequency of the ultrasonic/megasonic power supply is set to f _{1 first} , then f ₃ , and finally f ₄ , and f _{4 is} smaller than f ₃ , and f _{3 is} smaller than f ₁ .

Similar to the method illustrated in Fig. 9C, in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₄ first, then f ₃ , and finally f ₁ , and f _{4 is} smaller than f ₃ , and f _{3 is} smaller than f ₁ .

Similar to the method illustrated in FIG. 9C, in step 4, the frequency of the ultrasonic/megasonic power supply is set to f _{1 first} , then f ₄ , and finally f ₃ , and f _{4 is} smaller than f ₃ , and f _{3 is} smaller than f ₁ .

Similar to the method illustrated in FIG. 9C, in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₃ first, then f ₄ , and finally f ₁ , and f _{4 is} smaller than f ₃ , and f _{3 is} smaller than f ₁ .

Similar to the method illustrated in Fig. 9C, in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₃ first, then f ₁ , and finally f ₄ , and f _{4 is} smaller than f ₃ , and f _{3 is} smaller than f ₁ .

Similar to the method illustrated in Fig. 9C, in step 4, the frequency of the ultrasonic/megasonic power supply is set to f ₄ first, then f ₁ , and finally f ₃ , and f _{4 is} smaller than f ₃ , and f _{3 is} smaller than f ₁ .

10A-10B illustrate the use of ultrasonic/megasonic waves to clean the wafer according to the present invention to achieve zero damage to the patterned structure on the wafer by maintaining stable air cavity oscillation. Fig. 10A is the output waveform of the power supply, and Fig. 10B is the temperature curve corresponding to each cycle of cavitation oscillation. The operation process steps proposed by the present invention are as follows:

Step 1: Place the ultrasonic/megasonic device near the surface of the wafer or substrate set on the chuck or solution tank;

Step 2: Fill the space between the wafer and the ultrasonic/megasonic device with chemical liquid or gas mixed with water;

Step 3: Spin the chuck or vibrate the wafer;

Step 4: Set the power supply frequency to f ₁ and power to P ₁ ;

Step 5: Before the gas or steam temperature in the bubble reaches the implosion temperature T _i (total time τ ₁ elapses), set the power output frequency to f ₁ , power to P ₂ , and P _{2 to be} less than P ₁ . Therefore, since the temperature of the liquid or water is much lower than the temperature of the gas, the temperature of the gas inside the bubble begins to drop.

Step 6: The temperature of the gas in the bubble is reduced to close to the normal temperature T ₀ or the time (time of zero power) reaches τ ₂ , and the power supply frequency is set to f ₁ and the power to P _{1 again} .

Step 7: Repeat step 1 to step 6 until the wafer is cleaned.

In step 6, since the power is P ₂ , the temperature of the gas in the bubble cannot fall to room temperature, and a temperature difference ΔT _{2 needs to} exist in the time interval τ ₂ , as shown in Fig. 10B.

11A-11B illustrate a wafer cleaning method using an ultrasonic/megasonic device according to the present invention. It is similar to the method illustrated in FIG. 10A, except that step 5 sets the frequency of the ultrasonic/megasonic power supply to f ₂ and the power to P ₂ , where f _{2 is} smaller than f ₁ , and P _{2 is} smaller than P ₁ . Since f _{2 is} smaller than f ₁ , the temperature of the gas or steam in the bubble rises rapidly, so P ₂ should be much smaller than P _1. In order to reduce the temperature of the gas or steam in the bubble, the difference between the two is preferably 5 times or 10 times.

Figures 12A-12B illustrate a wafer cleaning method using an ultrasonic/megasonic device according to the present invention. It is similar to the method illustrated in FIG. 10A, except that step 5 sets the frequency of the ultrasonic/megasonic power supply to f ₂ and the power to P ₂ , where f _{2 is} greater than f ₁ , and P ₂ is equal to P ₁ .

Figures 13A-13B illustrate a wafer cleaning method using an ultrasonic/megasonic device according to the present invention. It is similar to the method illustrated in FIG. 10A, except that step 5 sets the frequency of the ultrasonic/megasonic power supply to f ₂ and the power to P ₂ , where f _{2 is} greater than f ₁ and P _{2 is} less than P ₁ .

14A-14B illustrate a wafer cleaning method using an ultrasonic/megasonic device according to the present invention. It is similar to the method illustrated in FIG. 10A, except that step 5 sets the frequency of the ultrasonic/megasonic power supply to f ₂ and the power to P ₂ , where f _{2 is} greater than f ₁ , and P _{2 is} greater than P ₁ . Since f _{2 is} greater than f ₁ , the temperature of the gas or vapor in the bubble rises slowly, so P ₂ can be slightly greater than P ₁ , but it must be ensured that the temperature of the gas or vapor in the bubble _{in the time interval τ 2} is more than the _{time interval τ 1} Decrease, as shown in Figure 14B.

Figures 4A-4B illustrate that the patterned structure is damaged by violent micro-jetting. 15A-15B illustrate that stable cavitation oscillation can also damage the patterned structure on the wafer 15010. As the cavitation oscillation continues, the gas in the bubble Or the steam temperature rises, so the size of the bubble 15046 is also increasing, as shown in Figure 15A. When the size of the bubble 15048 becomes larger than the spacing W between the patterned structures shown in FIG. 15B, the expansion of the air cavity oscillation will cause damage to the patterned structure 15034, as shown in FIG. 15C. The following is another cleaning method proposed by the present invention:

Step 1: Place the ultrasonic/megasonic device near the surface of the wafer or substrate set on the chuck or solution tank;

Step 2: Fill the space between the wafer and the ultrasonic/megasonic device with chemical liquid or gas mixed with water;

Step 3: Spin the chuck or vibrate the wafer;

Step 4: Set the power supply frequency to f ₁ and power to P ₁ ;

Step 5: Before the size of the bubble reaches the distance W between the patterned structures (time τ ₁ elapses), set the output power of the power supply to 0 watts. Since the temperature of the liquid or water is much lower than the gas temperature, the gas temperature in the bubble begin to drop.

Step 6: After the gas temperature in the bubble is cooled to normal temperature T ₀ or the time (time of zero power) reaches τ ₂ , set the power supply frequency to f ₁ and power to P _{1 again} .

Step 7: Repeat step 1 to step 6 until the wafer is cleaned.

In step 6, the temperature of the gas in the bubble does not have to drop to room temperature, it can be any temperature, but it is better to be much lower than the implosion temperature T _i . In step 5, the size of the bubble may be slightly larger than the size of the spacing between the patterned structures, as long as the expansion force of the bubble does not damage the patterned structure. The time τ ₁ can be determined by the following method:

Step 1: Similar to Table 1, select 5 different times τ ₁ as the conditions of the DOE experiment;

Step 2: selecting at least a period of time τ ₁ 10 times τ _2, the best choice for the initial test Zhan 100 times;

Step 3: Use the determined power P _{0 to} run the above five conditions to clean the wafers with patterned structures. Here, P ₀ is the patterning of the wafer in the continuous uninterrupted mode (non-pulse mode) The power at which the structure causes damage.

Step 4: Use SEMS or wafer pattern damage viewing tools to check the damage degree of the above five types of wafers, such as AMAT SEM view or Hitachi IS3000, and then the damage time τ ₁ can be determined within a certain range.

Repeat steps 1 to 4 to narrow the range of _{damage time τ d.} Knowing the damage time τ _d , τ ₁ can be set to a value _{less than 0.5τ d under the safety factor.}

All the methods described in FIGS. 7-14 are applicable to this or combined with the method described in FIG. 15.

Fig. 16 shows an embodiment of a wafer cleaning device using an ultrasonic/megasonic device. The wafer cleaning device includes a wafer 16010, a wafer chuck 16014 driven and rotated by a rotating drive device 16016, a nozzle 16064 spraying cleaning liquid chemicals or deionized water 16060, an ultrasonic/megasonic device 16062 combined with a nozzle 16064, and ultrasonic/ Megasonic power supply. The ultrasonic/megasonic wave generated by the ultrasonic/megasonic device 16062 is transmitted to the wafer through a liquid column of chemical reagents or deionized water 16060. All the cleaning methods described in FIGS. 7 to 15 are applicable to the cleaning device shown in FIG. 16.

Fig. 17 is an embodiment of a wafer cleaning device using an ultrasonic/megasonic device. Wafer cleaning device includes wafer 17010, solution tank 17074. A wafer cassette 17076 placed in the solution tank 17074 to support the wafer 17010, a cleaning solution chemical reagent 17070, an ultrasonic/megasonic device 17072 and an ultrasonic/megasonic power supply set on the outer wall of the solution tank 17074. At least one inlet is used to fill the cleaning solution chemical reagent 17070 into the solution tank 17074 to immerse the wafer 17010. All the cleaning methods described in FIGS. 7 to 15 are applicable to the cleaning device shown in FIG. 17.

18A-18C illustrate an embodiment of a method for cleaning a wafer using an ultrasonic/megasonic device according to the present invention. This method is similar to the method shown in Figure 7A, except that in step 5, before the gas or vapor temperature in the bubble reaches the implosion temperature T _i (or the time reaches τ ₁ <τ _i , τ _i is calculated by equation (11)), Set the output value of the power supply to a positive or negative DC value to maintain or stop the vibration of the ultrasonic/megasonic device. Therefore, because the temperature of the liquid or water is much lower than the gas temperature, the gas temperature in the bubble begins to drop. The positive or negative value here can be greater than, equal to, or less than the power P ₁ .

FIG. 19 illustrates an embodiment of a method for cleaning a wafer using an ultrasonic/megasonic device according to the present invention. Similar to the method shown in Figure 7A, except that in step 5, before the gas or vapor temperature in the bubble reaches the implosion temperature T _i (or the time reaches τ ₁ <τ ₂ , τ _i is calculated by equation (11)), the power supply is set The output frequency of is the same as _{f 1 , and the phase is opposite to that of f 1} to quickly stop the bubble’s cavitation oscillation. Therefore, since the temperature of the liquid or water is much lower than the temperature of the gas, the temperature of the gas inside the bubble begins to drop. The positive or negative value here can be greater than, equal to, or less than the power P ₁ . During the above operation, the output frequency of the power supply can be _{different from the frequency f 1} but the phase is opposite to the phase of f ₁ to quickly stop the cavitation oscillation of the bubbles.

Generally speaking, ultrasonic/megasonic waves with a frequency range of 0.1MHZ-10MHZ can be applied to the method proposed in the present invention.

Although the present invention is illustrated by specific embodiments, examples, and applications, obvious modifications and replacements in the art will still fall within the protection scope of the present invention.

1003‧‧‧Ultrasonic/Megasonic Device

1004‧‧‧Piezoelectric sensor

1008‧‧‧Acoustic Resonator

1010‧‧‧wafer

1012‧‧‧Nozzle

1014‧‧‧wafer chuck

1016‧‧‧Drive device

1032‧‧‧Deionized water (cleaning liquid chemical reagent)

Claims (64)

A method for cleaning a semiconductor substrate without damaging the patterned structure on the semiconductor substrate using a super/megasonic device, which is characterized in that it comprises: spraying liquid into the gap between the semiconductor substrate and the super/megasonic device; Set the frequency of the ultra/megasonic power supply to f ₁ and the power to P ₁ to drive the ultra/megasonic device; elapsed time τ ₁ , before the cavity oscillation in the liquid damages the patterned structure on the semiconductor substrate , Set the frequency of the super/megasonic power supply to f ₂ and the power to P ₂ to drive the super/megasonic device; elapsed time τ ₂ , after the temperature in the bubble cools to the set temperature, set the super The frequency of the megasonic power supply is f ₁ and the power is P ₁ , where τ ₂ >τ ₁ ; repeat the above steps until the semiconductor substrate is cleaned. The method according to claim 1, wherein the damage to the patterned structure on the semiconductor substrate by the cavitation oscillation is caused by micro-jets generated by bubble implosion. The method according to claim 1, wherein the frequency of the super/megasonic power supply is set to f ₁ and the power is P ₁ and the frequency of the super/megasonic power supply is set to f ₂ and the power is P ₂ The time interval between is less than 2000 times the waveform period _{of frequency f 1.} The method according to claim 1, wherein the frequency of the super/megasonic power supply is set to f ₁ and the power is P ₁ and the frequency of the super/megasonic power supply is set to f ₂ and the power is P ₂ The time interval between is less than ((T _i -T _0- △T)/(△T-δT)+1)/f ₁ , where T _i is the temperature of the gas and steam inside the bubble when the bubble imploses, T ₀ is the temperature of the liquid, ΔT is the temperature increase after the bubble is compressed once, and δT is the temperature decrease after the bubble is expanded once. The method according to claim 1, wherein the cavitation oscillation damages the patterned structure on the semiconductor substrate because the bubble size grows larger than the distance between the patterned structures. The method according to claim 1, wherein the set temperature is close to the temperature of the liquid. The method according to claim 1, wherein _{the value of the power P 2} is set to zero. _{The method according to claim 1} , wherein the frequency f 1 is equal to the frequency f ₂ , and the power P _{2 is} less than the power P ₁ . The method according to claim _{1, wherein the frequency f 1 is} higher than the frequency f ₂ , and the power P _{2 is} smaller than the power P ₁ . The method according to claim _{1, wherein the frequency f 1 is} less than the frequency f ₂ , and the power P ₁ is equal to the power P ₂ . The method according to claim _{1, wherein the frequency f 1 is} less than the frequency f ₂ , and the power P _{1 is} greater than the power P ₂ . The method according to claim _{1, wherein the frequency f 1 is} smaller than the frequency f ₂ , and the power P _{1 is} smaller than the power P ₂ . The method according to claim 1, wherein the output power P _{1 of} the super/megasonic power source has a gradually increasing amplitude. The method according to claim 1, wherein the output power P _{1 of} the super/megasonic power source has a gradually decreasing amplitude. The method according to claim 1, wherein the output power P _{1 of} the super/megasonic power source has an amplitude that first increases and then decreases. The method according to claim 1, wherein the output power P _{1 of} the super/megasonic power source has an amplitude that first decreases and then increases. The method according to claim 1, characterized in that the frequency of the _{output power P 1 of the super/megasonic power source is f 1 first and} then f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, wherein the frequency of the _{output power P 1 of the super/megasonic power source is f 3} first and then f ₁ , and f _{3 is} less than f ₁ . The method according to claim 1, characterized in that the frequency of the _{output power P 1 of the super/megasonic power source is f 3} first, then f _{1 and} finally f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, characterized in that, the frequency of the _{output power P 1 of the super/megasonic power source is first f 1,} then f _{3 and} finally f ₁ , and f _{3 is} less than f ₁ . The method according to claim 1, characterized in that the frequency of the _{output power P 1 of the super/megasonic power source is first f 1,} then f _{3 and} finally f ₄ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, characterized in that the frequency of the _{output power P 1 of the super/megasonic power supply is f 4} first, then f _{3 and} finally f ₁ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, characterized in that, the frequency of the _{output power P 1 of the super/megasonic power source is first f 1,} then f _{4, and} finally f ₃ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, characterized in that the frequency of the _{output power P 1 of the super/megasonic power supply is f 3} first, then f _{4 and} finally f ₁ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, wherein the frequency of the _{output power P 1 of the super/megasonic power source is f 3} first, then f _{1 and} finally f ₄ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, wherein the frequency of the _{output power P 1 of the super/megasonic power source is f 4} first, then f _{1 and} finally f ₃ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The method according to claim 1, wherein the frequency f ₂ is 0, and the power P ₂ is a positive value. The method according to claim 1, wherein the frequency f ₂ is 0, and the power P ₂ is a negative value. The method according to a request, wherein the frequency f ₂ is equal to _1, the phase of the phase f ₂ F ₁ opposite to f. The method according to a request, wherein the frequency f ₂ is different from the f _1, f ₂ opposite phase to the phase of the f _1. A device for cleaning a semiconductor substrate using a super/megasonic device, comprising: a chuck supporting the semiconductor substrate; a super/megasonic device placed near the semiconductor substrate; at least one spray head spraying chemical liquid to the semiconductor substrate and the semiconductor In the gap between the substrate and the ultra/megasonic device; the ultra/megasonic power supply; set the frequency of the ultra/megasonic power supply to f ₁ and the power to be P ₁ to drive the ultra/megasonic device; elapsed time τ ₁ Before the cavity oscillation in the liquid damages the patterned structure on the semiconductor substrate, set the frequency of the super/megasonic power supply to f ₂ and the power to P ₂ to drive the super/megasonic device; elapsed time τ _{2. After} the temperature in the bubble cools to the set temperature, set the frequency of the ultra/megasonic power supply to f _{1 again} and the power to P ₁ , where τ ₂ >τ ₁ ; repeat the above steps until the semiconductor substrate is cleaned. The device according to claim 31, characterized in that the damage to the patterned structure on the semiconductor substrate by the cavitation oscillation is caused by micro-jets generated by bubble implosion. The device according to claim 31, wherein the frequency of the super/megasonic power supply is set to f ₁ and the power is P ₁ and the frequency of the super/megasonic power supply is set to f ₂ and the power is P ₂ The time interval between is less than 2000 times the waveform period _{of frequency f 1.} The device according to claim 31, wherein the frequency of the super/megasonic power supply is set to f ₁ and the power is P ₁ and the frequency of the super/megasonic power supply is set to f ₂ and the power is P ₂ The time interval between is less than ((T _i -T ₀ -△T)/(△T-δT)+1)/f ₁ , where T _i is the temperature of the gas and steam inside the bubble when the bubble imploses, T ₀ is the temperature of the liquid, ΔT is the temperature increase after the bubble is compressed once, and δT is the temperature decrease after the bubble is expanded once. The device according to claim 31, characterized in that the damage to the patterned structure on the semiconductor substrate by the cavitation oscillation is caused by the increase in the size of the bubble larger than the spacing between the patterned structures. The device according to claim 31, wherein the set temperature is close to the temperature of the liquid. The device according to claim 31, wherein the value of _{the power P 2 is zero.} The device according to claim 31, wherein the frequency f ₁ is equal to the frequency f ₂ , and the power P _{2 is} less than the power P ₁ . The device according to claim 31, wherein the frequency f _{1 is} higher than the frequency f ₂ , and the power P _{2 is} smaller than the power P ₁ . The device according to claim 31, wherein the frequency f _{1 is} less than the frequency f ₂ , and the power P ₁ is equal to the power P ₂ . The device according to claim 31, wherein the frequency f _{1 is} less than the frequency f ₂ , and the power P _{1 is} greater than the power P ₂ . The device according to claim 31, wherein the frequency f _{1 is} smaller than the frequency f ₂ , and the power P _{1 is} smaller than the power P ₂ . The device according to claim 31, wherein the output power P _{1 of} the super/megasonic power source has a gradually increasing amplitude. The device according to claim 31, wherein the output power P _{1 of} the super/megasonic power source has a gradually decreasing amplitude. The device according to claim 31, wherein the output power P _{1 of} the super/megasonic power source has an amplitude that first increases and then decreases. The device according to claim 31, wherein the output power P _{1 of} the super/megasonic power source has an amplitude that first decreases and then increases. The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power source is f 1 first and} then f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power source is f 3} first and then f ₁ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power supply is f 3} first, then f _{1 and} finally f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, characterized in that the frequency of the _{output power P 1 of the super/megasonic power source is first f 1,} then f _{3 and} finally f ₁ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power supply is first f 1,} then f _{3 and} finally f ₄ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power source is f 4} first, then f _{3 and} finally f ₁ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power source is first f 1,} then f _{4 and} finally f ₃ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power supply is f 3} first, then f _{4 and} finally f ₁ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power supply is f 3} first, then f _{1 and} finally f ₄ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency of the _{output power P 1 of the super/megasonic power supply is f 4} first, then f _{1 and} finally f ₃ , f _{4 is} less than f ₃ , and f _{3 is} less than f ₁ . The device according to claim 31, wherein the frequency f ₂ is 0, and the power P ₂ is a positive value. The device according to claim 31, wherein the frequency f ₂ is 0, and the power P ₂ is a negative value. The apparatus of claim 31 requests, wherein said frequency f ₂ is equal to _1, the phase of the phase f ₂ F ₁ opposite to f. The apparatus of claim 31 requests, wherein said frequency f ₂ is different from the f _1, f ₂ opposite phase to the phase of the f _1. A device for cleaning a semiconductor substrate using an ultra/megasonic device, comprising: a box supporting at least one semiconductor substrate; a solution tank containing the box; an ultra/megasonic device arranged on the outer wall of the solution tank; at least one inlet Fill the solution tank with chemical liquid to immerse the semiconductor substrate; super/megasonic power supply; set the frequency of the super/megasonic power supply to f ₁ and power to P ₁ to drive the super/megasonic device; At time τ ₁ , before the cavity oscillation in the liquid damages the patterned structure on the semiconductor substrate, the frequency of the ultra/megasonic power supply is set to f ₂ , and the power is set to P ₂ to drive the ultra/megasonic device; After the time τ ₂ , after the temperature in the bubble cools to the set temperature, set the frequency of the ultra/megasonic power supply to f ₁ and the power to P ₁ , where τ ₂ >τ ₁ ; repeat the above steps until the semiconductor substrate is washed net. The device according to claim 61, wherein the power P ₂ is zero. A device for cleaning a semiconductor substrate using a super/megasonic device, comprising: a chuck supporting the semiconductor substrate; a super/megasonic device with a spray head placed near the semiconductor substrate, the spray head spraying chemicals to the semiconductor substrate Liquid; Ultra/Megasonic power supply; Set the frequency of the Ultra/Megasonic power supply to f ₁ and the power to P ₁ to drive the Ultra/Megasonic device; After time τ ₁ , the cavitation oscillation in the liquid damages the semiconductor Before the patterned structure on the substrate, set the frequency of the ultra/megasonic power supply to f ₂ and the power to P ₂ to drive the ultra/megasonic device; elapsed time τ ₂ , after the temperature in the bubble is cooled to the set temperature , Again set the frequency of the ultra/megasonic power supply to f ₁ and the power to P ₁ , where τ ₂ >τ ₁ ; repeat the above steps until the semiconductor substrate is cleaned. The device according to claim 63, wherein the power P ₂ is zero.

Images