JPS63164219A - Method and apparatus for monitoring depth - Google Patents

Abstract

PURPOSE:To monitor a working depth in high accuracy even if a beam current varies during working by measuring the current at a predetermined time interval, time-integrating the product of the current and a working speed coefficient to obtain a working volume, and dividing the volume by a beam scanning range area to obtain the depth. CONSTITUTION:A working machine with focused ion beam having an energy beam, such as the focused ion beam, a deflection scanning system for scanning the energy beam on a sample face, and a mechanism for blanking the beam is used. In such a case, the working volume (working speed coefficient) K of a workpiece per unit time by a unit current is obtained in advance, the beam current iB is measured at each predetermined time ts during working, or other physical amount for accurately calculating the beam current is measured, and the current iB is calculated therefrom. The product of the current is and the coefficient K is time-integrated to obtain a working volume V, which is divided by a beam scanning area A to obtain a working depth Z.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a processing depth monitoring technique when processing an LSI or an exposure mask using a focused ion beam or the like.

[Conventional technology]

In recent years, in the LSI development process, it has become very important to debug, correct, or analyze defective parts by cutting or connecting part of the wiring within the LSI chip. For this purpose, examples have been reported in which LSI wiring is cut using a focused ion beam.

For example, in Japanese Patent Application Laid-Open No. 58-106750 (Focused Ion Beam Processing Method), the ion beam irradiation amount is disclosed.

It has been shown that processing with different etch depths is possible by changing the irradiation time, acceleration voltage, etc.

[Problem that the invention seeks to solve]

In the above-mentioned conventional technology, the only way to determine when the desired processing depth has been obtained and to stop processing is to vary the irradiation time, irradiation amount, etc. . Although a formula representing the etching depth S is shown in P4, there is no specific description of how to achieve the target depth using this formula. As for a specific method of detecting the end point of machining, a method of analyzing secondary ions is mentioned.

However, LSI uses multilayer wiring, and in order to cut the lower layer wiring, as shown in the cross section of the processed part in Figure 3,
Typically, it is necessary to drill holes with a high aspect ratio, such as a processing area of 5 μm and a processing depth of 10 μm. When the machining depth is shallow as shown in Fig. 2, sufficient secondary ions 20 are captured by the secondary ion detector 21, but when the aspect ratio becomes high as shown in Fig. 4, Ion 20 was almost detected (because of this,
It is impossible to detect the machining end point with this method.

On the other hand, if the beam current and accelerating voltage are constant, the machining depth is proportional to the machining time. Since the machining time is short when the machining depth is shallow, we assumed that the beam current was constant within this time and controlled the depth using the machining time without causing a large error.

However, for the hole shown in Figure 3, for example, at a typical machining speed: 0.14 5/s, the volume: 5X5X10=25
It takes about 30 minutes to process 0i, and within this time the drift of the beam current iB cannot be ignored as shown in FIG. Drift can exceed 10%.

Therefore, as shown in Figure 5, if the processing time is set based on the initially set current value and then the beam current drifts in a decreasing direction, the actual depth will be insufficient and the wiring 22 will not be cut. Can not. On the other hand, if the actual current value drifts in the direction of increasing from the initially set current value, the processing will be deeper than the target depth, and the lower layer wiring will also be processed, resulting in re-spatter adhesion from the lower layer wiring. This causes problems such as short circuits with upper layer wiring.

An object of the present invention is to monitor depth with high accuracy even if the beam current changes during processing.

[Means for solving problems]

The above purpose is to measure the beam current at sufficiently short time intervals while machining one hole, time-integrate the product of this and the machining speed coefficient to find the machining volume, and divide this by the area of the beam scan area for machining. This is achieved by gaining depth.

[Effect]

In processing using an ion beam or the like, the sample is shaved by sputtering, so as the hole becomes deeper, the effect of re-adhering the sputtered particles 25 to the side wall increases as the hole becomes deeper, and as a result, the hole becomes tapered. . Therefore, the opening has the same area A as the beam scanning area, but as the depth increases, the bottom area A' becomes A'<A.

As a result of the experiment, as shown in FIG. 7, the increase rate of the machined hole volume decreases with respect to the processing time (the beam current can be considered to be approximately constant). However, as shown in FIG. 8, the machining depth increases at a constant rate as long as a flat portion remains on the bottom (A>0). However, it was found that this relationship breaks down when the bottom surface disappears (A'=0) and the hole becomes conical.

From this, as shown in Fig. 7, the volume V of the material sputtered by the beam is constant per time, but the redeposition volume V2 increases as the hole becomes deeper, so the large processing volume vl = v- It is considered that the increase rate of v2 changes.Here, the volume V of the material sputtered by the beam is distinguished from the machined hole volume v2, and is called the machined volume ■.

The machining volume V is: V = /kiBdt However: Machining speed coefficient (RL"A-'me-' ]
iB: Beam current [A]. Since this volume corresponds to the volume of the processed hole when there is no redeposition of sputtered particles, the non-tapered cross-sectional area becomes the volume of a rectangular prism of A (beam scan area area) everywhere.

Therefore, the trap to find machining depth 2 is Z = V/A And it is sufficient.

〔Example〕

A first embodiment of the present invention will be described below with reference to FIG. 1 and FIGS. 9 to 11.

In FIG. 7, the ion beam extracted from the ion source 1 is connected to the first ion beam. Second. Third lens electrode (respectively 2 in the figure)
．． 3.4), the light is focused on the sample 8 to bring it into focus. By applying a voltage to the planking electrode 5 as necessary, the beam can be bent to hit the planking aperture 6, thereby eliminating irradiation of the sample 8. By applying a deflection voltage to the deflector electrode 7, the beam can be scanned within the processing area.

The sample 8 is fixed to a stage 9, and the stage 9 is driven by a drive device (not shown). Stage 9 during processing
is fixed and the beam is deflected and processed using a def-reft tough.

This device includes a deflector controller 10. Planking controller 11. Acceleration power source 12. Output power supply 13
This provides the necessary voltage.

FIG. 1 shows a flow of monitoring the machining depth using the apparatus shown in FIG. 9. At the same time as the start of the injection, a timer is activated to measure the beam current iB at fixed time intervals t. time t
, is selected at a time when beam current fluctuations in this range can be ignored. When measuring the beam current, the stage 9 is moved from the processing position as shown in FIG. 8 so that the beam falls into the Faraday cup 19. Planking is applied to the beam while the stage 9 is moving, and the timer is not activated during the moving time of the stage 9 and the measurement time by the Faraday cup 19 because no processing is being performed. Combiator 19
The irradiation amount W, machining volume V, and machining depth 2 are determined by the following formula.

W=ΣiBt, [A-Ki] v=Ni-WCIMLI] Z=V/ACμm] However, k: Machining speed coefficient [□”A-”se-”]
A: Beam scan area [μ-] Processing speed coefficient includes sample material and ion energy.

It is determined by the ionic material, but in normal processing, the ion energy and ionic material are constant. Therefore, when machining a single material, the machining speed coefficient Kt may be treated as a constant number. According to experiments, in the case of ion energy 20KV and ionic material Ga, KSiOz = 0.2 for 5i02.
If the sample has a multilayer structure made of multiple materials, it must be taken into consideration that the machining speed coefficient K differs depending on the material.

This embodiment will be explained with reference to FIG.

! As shown in Figure 11, based on the depth 2 at the time of measuring the beam current iB, the material M currently being processed is determined from the thickness of each layer of the sample measured in advance.
The current machining volume ■ is determined by adding the machining volume increment Δ■ ΔV = kiltg (tun”) where k = k@ to the machining volume at the time of the previous beam current measurement. By dividing the volume V by the beam scanning area A, Z = V/A (μm) obtains the current depth Z. This depth Z is the target depth Z.

Processing will stop when it reaches .

In the first embodiment, since the stage 9 is moved every sampling time t, processing time is wasted.

A second embodiment that solves this problem will be explained with reference to FIGS. 12 to 14. In FIG. 12, ammeter 14 is an ammeter for measuring the source current is entering the third lens electrode 2. In FIG. The beam current iB is expressed by the function in = f(is) of the source current is as shown in FIG.
Put it in. Normally, sufficient accuracy can be obtained from this function using a -order function iB=ais+β.

Since the ammeter 14 is floating by the amount of accelerating voltage with respect to the ground, this analog measurement value is converted into a digital value by the A/D converter 15, and then coupled by the optical isolator 16 and sent to the computer 17 at the ground level. Enter.

In the flowchart of FIG. 1 or FIG. 11, the second embodiment can be achieved by measuring the source current iB and calculating the beam current 1n=f(is) from this value instead of measuring the beam current iB. can represent the flow of

FIG. 14 shows the results of an experiment conducted using this example.

In this experiment, the machining speed was approximately 0.3 S, and the machined hole was 5μ.
Since m0, the machining depth is 8 μmf? : The time to obtain is about 11 minutes. During this time, the beam current drifts, and when using the conventional technique of controlling the machining time based on the beam current value at the start of machining, the deviation of the machining depth from the target depth is ±1 μm.

On the other hand, according to the method of the present invention, the sampling time is 20
When the source current rB was measured in seconds, the machining depth deviation was reduced to ±0.25 μm. Since the thickness of the insulating layer between nine wiring layers of a normal LSI is about 1 μm, sufficient accuracy can be obtained by using the present invention.

Further, a third embodiment of the present invention will be explained with reference to FIGS. 15 and 16. In FIG. 15, the ammeter 18 measures the aperture current iA flowing into the third pole (beam limiting aperture) 4. In FIG. The beam current iB is
Since it can be expressed as a function of the aperture current iA, in = g(iA), the machining depth can be monitored in the same manner as in the second embodiment. This function also iB = aiA
It can be expressed with sufficient accuracy by a −order function of +β.

It is also possible to measure the 1 flow iB flowing into the ion source 1 in FIG. 12 at section A. If the electrode 2 has a shape that wraps in from above as shown in FIG. 12, the generation of secondary electrons due to ion irradiation can be suppressed, so the position of the ammeter 14 can also be adjusted.
Although an accurate source current can be measured, if the electrode 2 has a flat plate shape, secondary electrons will be generated by ions, and the ammeter 14 will measure a current larger than the inflowing ion current.

In this case, it is preferable to place an ammeter at section A in Figure 12.

Further, an embodiment of claim 3 will be explained with reference to FIG. 17. Measure the source current is in the same way as in FIG.
D converter 15. Optical isolator 16. D/A converter 26
Obtain the earth-level analog signal is. This is made into a beam current value iB by the addition multiplication circuit 27, and the multiplication circuit 28. The processing volume V is obtained by the integrating circuit 29. This is divided by the beam scanning area A by the multiplier circuit 30 and the depth Z
and is displayed on the display 31.

Furthermore, the comparison circuit 32 compares the depth Z and the target depth Zo, and outputs a machining end point signal when 2≧Zo. This signal causes the planking controller 11 to plank the beam and finish the machining.

〔Effect of the invention〕

According to the present invention, even when machining is performed over a period of time where fluctuations in beam current cannot be ignored, the machining depth can be monitored based on current values measured at sufficiently short time intervals, so holes can be machined with high depth accuracy. There is an effect that can be done.

[Brief explanation of the drawing]

FIG. 1 is a flowchart of the first embodiment of the present invention;
3 and 3 are cross-sectional views for explaining the secondary ion detection method, FIG. 4 is a diagram showing the time change of the beam current, FIG. 5 is a diagram showing the processing state by the conventional method, and FIG. 6 is the processing A cross-sectional view of the hole, FIGS. 7 and 8 are graphs showing the experimental results of machining volume and depth, and FIG. 9 is an apparatus configuration diagram showing the first embodiment.
Fig. 10 is a beam current measurement diagram in the first embodiment, Fig. 11 is a flowchart for processing multiple materials in the first embodiment, Fig. 12 is an apparatus configuration diagram showing the second embodiment, and Fig. 13 The figure is a graph showing the relationship between source current and beam current, and FIG. 14 is a graph showing experimental results according to the second embodiment.
Fig. 15 is an apparatus configuration diagram showing the third embodiment, Fig. 16 is a graph showing the relationship between aperture current and beam current, and Fig. 16 is a graph showing the relationship between aperture current and beam current.
FIG. 7 is an apparatus configuration diagram showing a fourth embodiment. 1... Ion source, 2... Second lens electrode,
3... Second lens electrode, 4... Third lens electrode (beam limiting aperture), 5... Planking electrode, 6... Planking aperture, 7... Deflector electrode, 8... - Sample, 9... Stage, 10... Deflector controller, 11... Planking controller, 12... Acceleration power source, 13... Extraction power source, 14, 18,
20...Ammeter, 15...A/D converter, 16
...Optical isolator, 17...CPU. 19... Faraday cup, 20... Secondary ion, 21... Secondary ion detector, 22... Wiring, 23... Interlayer insulating film, 24...
・Devotion, 25...Spacta particles, 26・
...D/A converter, 27...addition/multiplication circuit, 28.
...Multiplication circuit, 29...Integrator circuit, 30...
・Multiplication circuit, 31... Display, 32... Comparison circuit, ■... Machining volume, vl... Machining hole volume, v2... Redeposition volume, Z... Machining depth,
Zo...Target machining depth, A...Beam scanning area, A'...Area of the bottom surface of the machined hole, K...Machining speed coefficient, KM...Machining speed coefficient for material M, iB... Beam current, iB...source current, iA...flow through aperture 1, t...current sampling time. 7・'1～゛, Patent Attorney Katsuo Ogawa 5,5 Figure 7 Figure 9 Figure 10 Prisoner Figure 11

Claims (1)

[Claims] 1. Processing using a processing device such as a focused ion beam that has an energy beam such as a focused ion beam, a deflection scanning system that scans the beam on a sample surface, and a mechanism for planking this beam. In this case, (a) the volume of the workpiece processed per unit time by unit current (hereinafter referred to as the processing speed coefficient) is determined in advance, (b) the beam current is measured at predetermined intervals during processing, or Measure other physical quantities that can accurately calculate the beam current, calculate the beam current from this, (c) time-integrate the product of this beam current and the machining speed coefficient to find the machining volume, and divide this by the area of the beam scan region. A depth monitoring method in energy beam machining such as a focused ion beam, characterized in that the machining depth is obtained by 2. In the depth monitoring method according to claim 1, when the workpiece is a multilayer sample made of a plurality of materials, (d) machining speed coefficient of each material and layer of each material; (e) Find the beam current using the method in (b), (f) Time-integrate the product of this beam current and the machining speed coefficient according to the material currently being machined to calculate the machining volume. A method for monitoring depth in machining using an energy beam such as a focused ion beam, characterized in that the machining depth is obtained by determining the machining depth by dividing the machining depth by the area of the beam scanning region. 3. Depth consisting of a device that measures the source current, etc., a circuit that calculates the beam current value from the source current value, a circuit that calculates the machining volume and machining depth from the beam current, and a device that displays the machining depth. Monitor device.