Cross thinning method for preparing TEM sample slice by using FIB

Abstract

The invention relates to the technical field of semiconductor manufacturing, provides a cross thinning method for preparing a TEM (Transmission Electron Microscope) sample sheet by utilizing FIB (Forward Ion Beam), and aims to solve the problem that bending deformation occurs due to insufficient material strength to support internal stress when the thickness of the TEM sample sheet is lower than a certain critical value in the prior art. The core of the method is that a cross thinning strategy is adopted, and in each stage of an FIB thinning process, asymmetric thinning of two sides of a sample is realized by setting different thinning lengths of a front slice and a rear slice and sequentially adjusting the length difference along with reduction of an ion beam current. The method has the technical effects that the accumulated stress in the sample can be effectively counteracted and released, so that the bending deformation of the ultra-thin (less than 50 nm) TEM sample is remarkably inhibited or completely avoided when the ultra-thin (less than 50 nm) TEM sample is finally obtained, the success rate and the quality of sample preparation are improved, and a guarantee is provided for subsequent accurate TEM observation and analysis.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing technology, and in particular to a method for preparing ultrathin TEM samples by using a cross-thinning strategy to prevent bending deformation. Background Technology

[0002] In existing standard procedures for preparing TEM samples using FIB (Film Injection Bypass), a symmetrical thinning method is typically employed. That is, in each thinning step (e.g., from roughing to finishing, with the beam current decreasing sequentially), the front and back surfaces of the sample are milled with ion beams of equal length and dose to balance processing efficiency and final surface quality. This symmetrical thinning method is effective for thicker samples, but it has inherent technical defects and limitations when preparing ultrathin TEM samples.

[0003] As the feature size of semiconductor devices continues to shrink, the thickness requirements for TEM samples are becoming increasingly stringent. When a sample is thinned to a critical thickness, its mechanical strength drops sharply. At this point, the non-uniform stresses generated within the material during FIB processing (such as lattice damage caused by gallium ion implantation and amorphous layer stress) cannot be balanced by its own structural strength. In symmetrical thinning mode, these stresses accumulate uniformly in the central plane of the sample. When the stress exceeds the yield limit of the material, it causes macroscopic and irreversible bending or curling deformation of the entire sheet. This deformation makes it impossible for the sample to remain flat in the TEM, resulting in the observation area deviating from the electron beam focus, blurred images, and even the inability to perform atomic-scale structural analysis, thus leading to fabrication failure. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method for preparing ultrathin TEM samples by employing a cross-thinning strategy to prevent bending deformation. The technical concept is as follows: The porous sample is fixed to the nail stage with conductive C glue, and then the sample is bonded to the nail stage with conductive Cu glue. The conductive Cu glue wraps the sample, leaving the area to be processed.

[0005] The prepared sample was fixed on a 38° sample stage and injected. First, a tungsten protective layer was deposited at the desired processing location using an ion beam. The sample was then etched using a conventional FIB process. An initial wedge-shaped sample was formed using a U-shaped cutting method. Combined with an in-situ nanomanipulator, the wedge-shaped sample was extracted to form the required observation area. During the etching process, the etching effect was optimized and sample damage reduced by adjusting FIB parameters (such as accelerating voltage and ion beam current).

[0006] For wedge-shaped samples initially formed by cutting, a U-shaped cutting tilt angle method is used. By adjusting the cutting angle of the ion beam, the wedge shape of the sample is gradually reduced. At the same time, combined with an in-situ nanomanipulator and a gas injection system (GIS), the cutting and thinning process of the sample is precisely controlled. During the thinning process, a cross-cutting and asymmetric method is used to thin the sample, achieving precise processing of the target sample.

[0007] The prepared TEM samples underwent post-treatment processes such as purging and cleaning to remove the amorphous and damaged layers from the surface. Simultaneously, the samples were further characterized and analyzed to ensure their quality and performance met the requirements.

[0008] Based on the above overall technical concept, this invention provides a cross-thinning method for preparing transmission electron microscopy (TEM) sample slices using focused ion beam (FIB), comprising the following steps: (a) Sample preparation: After fixing the sample with conductive C glue, it is then sprayed with gold to enhance conductivity; (b) Initial cutting: In the FIB device, a protective layer is deposited on the area of ​​the sample to be observed, and the sample is initially cut by precisely controlling the focused ion beam to form a wedge-shaped sample; (c) The wedge-shaped sample is extracted and transferred to a special sample carrier using an in-situ manipulation device; (d) The transferred wedge-shaped sample is subjected to cross-thinning treatment, wherein in at least one thinning step, the thinning length of the front slice of the sample is set to be different from the thinning length of the back slice, so as to achieve asymmetric cross-thinning of the sample. (e) The cross-thinned sample is post-processed to remove the surface amorphous layer and damage layer, complete the cross-thinning, and obtain a TEM sample sheet with a thickness of less than 50 nm.

[0009] Step (a) involves fixing the sample with conductive C glue, which includes fixing the sample on the sample stage with conductive carbon glue and reinforcing and wrapping it with conductive copper glue, leaving the area to be processed. Sputtering with gold refers to the process of sputtering gold onto a sample using an ion sputtering apparatus to enhance its conductivity. Finally, the nail stage for the loaded specimen and the copper mesh with the semi-circular copper ring were fixed on the flat stage and the 38° inclined stage of the FIB, respectively.

[0010] The initial cutting in step (b) includes: first, tilting the sample stage to 52°, depositing a tungsten protective layer at the target position of the sample, adjusting the ion beam voltage to 20-30 kV, and setting the ion beam current to 80 pA-0.79 nA; Then, a large beam current of 30-35kV and 9.3-27 nA was selected for the ion beam voltage and current, respectively. Rough pits were dug on the upper and lower sides and the left side of the tungsten coating using the regular cross section RCS mode. After rough processing, the upper and lower edges of the coating were finely processed using the cleaning cross section CCS mode at sample stage tilt angles of 50° and 54°, respectively, to form a preliminary sample sheet. Finally, after the sample stage returns to 0°, the thin sheet is U-cut to process it into a cantilever beam state.

[0011] The cross-thinning process in step (d) comprises multiple thinning steps performed sequentially, with the ion beam current used in each step decreasing sequentially.

[0012] In the plurality of sequentially performed thinning steps, the difference between the thinning length of the preceding slice and the thinning length of the subsequent slice decreases sequentially as the ion beam current decreases.

[0013] In step (d), the thinning length of the front slice is greater than the thinning length of the rear slice.

[0014] In step (d), during the thinning process, the tilt angle of the sample stage is set to ±1° to ±5°, the ion beam accelerating voltage is 2 kV to 30 kV, and the beam current range is 9 pA to 0.23 nA.

[0015] The post-processing described in step (e) includes: purging and cleaning the front and rear surfaces of the sample by gradually reducing the ion beam voltage and beam current to remove the amorphous layer and the damaged layer.

[0016] Another technical solution of the present invention is to provide a TEM sample sheet prepared by the method, wherein the thickness of the sample sheet is less than 50 nm and its bending deformation is significantly lower than that of a sample of the same thickness prepared by a symmetrical thinning method.

[0017] The present invention also provides a FIB device configured to perform the cross-thinning method as described above, wherein asymmetric thinning parameters are set by control software to achieve the cross-thinning.

[0018] This invention aims to provide a FIB cross-thinning method that can effectively prevent bending deformation of ultrathin TEM samples, overcoming the shortcomings of existing symmetrical thinning techniques. This method introduces asymmetric ion beam thinning parameters to actively release internal stress in the sample, preventing bending deformation at critical thicknesses.

[0019] Compared with existing technologies, the advantages of this patent are: 1. Effectively suppresses bending: By using the FIB cross-thinning method, the technical problem of easy bending and deformation of ultrathin TEM samples is fundamentally solved, and the success rate of ultrathin sample preparation is significantly improved.

[0020] 2. Improved observation quality: The prepared flat sample ensures that high-quality, high-resolution images can be obtained during TEM observation, which is especially important for the analysis of advanced semiconductor devices that require atomic-level resolution.

[0021] 3. Strong process compatibility: This invention does not require any hardware modification to existing FIB equipment. It can be achieved simply by optimizing the thinning parameter settings in the software, and is easy to integrate into existing semiconductor failure analysis and structural characterization processes.

[0022] 4. High cost-effectiveness: By reducing preparation failures and sample re-reproduction caused by sample bending, it saves expensive FIB machine time and labor costs, and accelerates the research and development progress and analysis cycle. Attached Figure Description

[0023] Figure 1 This is a schematic diagram showing the thinning length of the TEM sample slices before and after preparation in Example 1.

[0024] Figure 2 This is an example image of a TEM sample sheet prepared by the cross-thinning method in Example 1.

[0025] Figure 3 This is a schematic diagram of the bending deformation of the TEM sample prepared by symmetrical thinning in Example 2. Detailed Implementation

[0026] Example 1 Experiments were conducted on a Thermos Helios 5 UC-type FIB equipped with an in-situ nanomanipulator and a gas injection system (GIS). The invention is described in detail below with reference to specific embodiments: 1. Sample Preparation: Fix the sample to the sample stage using conductive C adhesive to ensure its stability. Next, use conductive Cu adhesive to bond the sample to the sample stage, forming a stable support structure. Then, perform gold sputtering on the sample using an ion sputtering instrument to enhance conductivity. Finally, fix the sample-loaded stage and the copper mesh with semi-circular copper rings on the FIB's flat stage and 38° inclined stage, respectively.

[0027] 2. Initial Cutting: In the FIB equipment, the sample is initially cut by precisely controlling the focused ion beam. First, the sample stage is tilted to 52°, and a 10μm × 2μm × 3μm tungsten protective layer is deposited at the target location on the sample. The ion beam voltage is adjusted to 30 kV. During this process, the ion beam current is set to a low value (e.g., 80 pA - 0.79 nA) to reduce damage to the sample surface. Then, a high ion beam voltage and current of 30 kV and 9.3-27 nA are selected, and a regular cross-section (RCS) pattern is used to rough-cut the tungsten coating on the upper, lower, and left sides. The rectangular frame should not be too close to the tungsten coating to avoid back deposition with the high beam current and ensure cutting accuracy. After rough cutting, the upper and lower edges of the coating are refined using a clean cross-section (CCS) pattern at sample stage tilt angles of 50° and 54°, respectively, to form a preliminary sample sheet. Finally, after the sample stage returned to 0°, the 2μm thick sheet was U-cut to process the sample sheet into a cantilever beam state.

[0028] 3. Sample Transfer: Adjust the relative positions of the nanorobotics and the sample sheet, use ion beam tungsten deposition for bonding, then cut the cantilever arm to quickly sever and separate the sample, extracting the sample. Adjust the tilt angle of the sample stage to 52°, with the copper column facing downwards. Insert the nanorobotics, readjust the relative positions of the nanorobotics and the copper column again, use a tungsten layer to bond the right side of the sample to the left end of the copper column, then cut the tungsten needle. Adjust the sample stage orientation back to 0°, with the copper column facing upwards.

[0029] 4. Cross-thinning: During the thinning process, the thinning voltage and ion beam current are 30 kV and 0.23 nA, respectively. The tilt angle is set to ±1°. Thinning is performed on both the front and back sides of the sample. The thinning length of the front slice is set to 8 μm, and the thinning length of the back slice is set to 6 μm, i.e., the difference in thinning length between the front and back slices is 2 μm (e.g., ...). Figure 1 The ion beam current was further reduced to 80 pA for cross-thinning. Asymmetric thinning parameters were set, and the difference in thinning length between the front and rear slices was gradually reduced from 2 μm to 1.5 μm, 1 μm, and 0.5 μm to ensure the integrity and undeformation of the area to be observed. The sample was finally thinned to the target thickness (30-50 nm). To further improve the sample quality and observation effect, the sample was cleaned by gradually reducing the ion beam voltage. First, the sample stage was tilted to ±3°, and the ion beam voltage and current were reduced to 5 kV and 15 pA, respectively. The front and rear surfaces of the sample were purged for 30 s each, 2-3 times. Then, the sample stage was tilted to ±5°, and the ion beam voltage and current were reduced to 2 kV and 9 pA, respectively. The front and rear surfaces of the sample were purged for 1 min each, stopping as needed to ensure that the tungsten layer remained. This step removes the amorphous layer and damaged layer on the sample surface, preserving the original structure and properties of the sample.

[0030] 5. Final Testing: Perform final testing on the prepared TEM samples. TEM observation ensures the samples meet the testing requirements. If necessary, further post-processing or optimization can be performed to improve observation results.

[0031] Example 2 The method is the same as in Example 1, except that the cross-thinning process is replaced with a symmetrical thinning process, wherein the symmetrical thinning process is as follows: After the sample is soldered onto the copper pillar, T is returned to 0°, and R is returned to 0°. At this point, the ion beam is 90° perpendicular to the sample. Then, CCS is used for thinning. This step reduces the overall width by 1 / 4 on both the top and bottom. Generally, T is rotated 1° to cut the front side and -1° to cut the back side, one cut on the front and one cut on the back, repeating the thinning process, with each cut slightly inward. The beam current is reduced, and the sample is focused and de-spread. During this process, the electron beam can be used for real-time observation. Generally, three cuts are taken per scan, continuing until the sample is thinned to 5kV transmittance under the electron beam. The beam current is continuously reduced until it reaches approximately 80pA. At this point, the vertical angle T is approximately ±0.8°. Cuts are made vertically to observe the position of the protective layer. The magnification at this point is approximately 10K Mag, and the thickness is approximately 100nm or less. Finally, fine-tuning is performed, with real-time observation using the ion beam.

Claims

1. A method for cross-thinning thin sections of transmission electron microscopy (TEM) samples using focused ion beam (FIB), characterized in that, Includes the following steps: (a) Sample preparation: After fixing the sample with conductive C glue, it is then sprayed with gold to enhance conductivity; (b) Initial cutting: In the FIB device, a protective layer is deposited on the area of ​​the sample to be observed, and the sample is initially cut by precisely controlling the focused ion beam to form a wedge-shaped sample; (c) The wedge-shaped sample is extracted and transferred to a special sample carrier using an in-situ manipulation device; (d) The transferred wedge-shaped sample is subjected to cross-thinning treatment, wherein in at least one thinning step, the thinning length of the front slice of the sample is set to be different from the thinning length of the back slice, so as to achieve asymmetric cross-thinning of the sample. (e) Post-process the cross-thinned sample to remove the surface amorphous layer and damage layer, thus completing the thinning.

2. The method according to claim 1, characterized in that, Step (a) involves fixing the sample with conductive C glue, which includes fixing the sample on the sample stage with conductive carbon glue and reinforcing and wrapping it with conductive copper glue, leaving the area to be processed. Sputtering with gold refers to the process of sputtering gold onto a sample using an ion sputtering apparatus to enhance its conductivity. Finally, the nail stage for the loaded specimen and the copper mesh with the semi-circular copper ring were fixed on the flat stage and the 38° inclined stage of the FIB, respectively.

3. The method according to claim 1, characterized in that, The initial cut described in step (b) includes: First, tilt the sample stage to 52°, deposit a tungsten protective layer at the target position of the sample, adjust the ion beam voltage to 20-30kV, and set the ion beam current to 80 pA-0.79 nA. Then, a large beam current of 30-35kV and 9.3-27 nA was selected for the ion beam voltage and current, respectively. Rough pits were dug on the upper and lower sides and the left side of the tungsten coating using the regular cross section RCS mode. After rough processing, the upper and lower edges of the coating were finely processed using the cleaning cross section CCS mode at sample stage tilt angles of 50° and 54°, respectively, to form a preliminary sample sheet. Finally, after the sample stage returns to 0°, the thin sheet is U-cut to process it into a cantilever beam state.

4. The method according to claim 1, characterized in that, The cross-thinning process in step (d) comprises multiple thinning steps performed sequentially, with the ion beam current used in each step decreasing sequentially.

5. The method according to claim 4, characterized in that, In the plurality of sequentially performed thinning steps, the difference between the thinning length of the preceding slice and the thinning length of the subsequent slice decreases sequentially as the ion beam current decreases.

6. The method according to claim 5, characterized in that, In step (d), the thinning length of the front slice is greater than the thinning length of the rear slice.

7. The method according to claim 1, characterized in that, In step (d), during the thinning process, the tilt angle of the sample stage is set to ±1° to ±5°, the ion beam accelerating voltage is 2 kV to 30 kV, and the beam current range is 9 pA to 0.23 nA.

8. The method according to claim 1, characterized in that, The post-processing described in step (e) includes: purging and cleaning the front and rear surfaces of the sample by gradually reducing the ion beam voltage and beam current to remove the amorphous layer and the damaged layer.

9. A TEM sample sheet prepared by the method according to any one of claims 1 to 8, characterized in that, The thickness of the sample sheet is less than 50 nm.

Images