A method for controlling stress fluctuations in a silicon oxycarbide-containing layer formed using plasma-enhanced chemical deposition.

Abstract

This system controls stress fluctuations in silicon oxycarbide-containing layers formed using a capacitively coupled plasma-enhanced chemical deposition system. [Solution] The method provides a chamber equipped with a platen for supporting a substrate, places a substrate on the platen on which a layer containing silicon oxycarbide can be formed, introduces an oxygen-containing gas into the chamber, introduces trimethylsilane gas into the chamber, generates a plasma in the chamber, maintains the temperature at which deposition occurs below 275°C, maintains the pressure in the chamber in the range of 1.5 to 3 Torr, and deposits a layer on the substrate at a deposition rate > 1.5 μm / min.

Description

Technical Field

[0001] The present invention relates to a method for controlling stress variations in a layer containing silicon oxycarbide formed using plasma chemical vapor deposition (plasma CVD (Chemical Vapour Deposition)).

Background Art

[0002] Dielectric thin films are widely used in the manufacture of semiconductor devices, MEMS devices, and optical and electro-optical devices. For example, stress is incorporated into the dielectric thin film during deposition processes such as chemical vapor deposition (CVD), plasma-enhanced chemical vapor deposition (PECVD), and atomic layer deposition (ALD). This stress can be caused by exogenous factors such as the difference in the coefficient of thermal expansion between the dielectric thin film and the underlying substrate, or endogenous factors such as defects in the microstructure of the dielectric thin film. Compressive or tensile stress that may exist within the dielectric thin film can lead to undesirable wafer distortion, which can then be a problem for subsequent process steps, or can cause failure in the dielectric thin film itself in the form of cracks or delamination.

[0003] As coating thickness increases, stress control becomes increasingly difficult. In applications requiring thick layers of SiO2 exceeding 10 μm, conventional PECVD processes tend to be preferred techniques for high deposition rate processes. These conventional processes tend to require low deposition rates or multiple stress equilibrium layers to achieve crack-free layers with minimal wafer strain. This results in complex deposition processes with low total deposition rates. For example, Patent Document 1 (US9472610B2) discloses that it is impossible to deposit SiO2 layers with a thickness of 20 μm or more at temperatures below 280°C without cracks due to stress accumulation within the layer, and therefore, alternating tensile and compressive stress lamination is used to form the SiO2 deposit. Patent Document 2 (US2012 / 015113 A1) discloses the formation of multiple layers of SiO2 using thermal quasi-atmospheric CVD (SACVD) and PECVD with a total layer thickness of up to 3.5 μm. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] US Patent 9472610 B2 Publication [Patent Document 2] U.S. Patent Application Publication 2012 / 015113 A1 [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] Since deposition rate, layer density, and stress are all affected by the deposition temperature, the thermal budget (the amount of heat the wafer can accept) is another constraint to consider when thinking about the deposition process. Many MEMS (Micro Electro Mechanical Systems) and 3D packaging applications have the constraint of requiring low temperatures, limiting the deposition temperature to below approximately 300°C to prevent device damage or warping.

[0006] In many schemes, wafer warpage after dielectric deposition needs to be limited to less than 300 μm across the wafer diameter. Achieving this warping limitation requires low stress fluctuations across the wafer diameter, which is typically facilitated by using low-speed dielectric deposition. However, to make the deposition process cost-effective, high deposition rates of 1.5 μm / min (>1.5 μm / min) or higher are required. Since high deposition rates typically increase stress within the dielectric layer, high deposition rates can lead to further challenges.

[0007] Current preferred dielectric materials for filling the gaps between dies on a wafer include tetraethyl orthosilicate (TEOS) oxides. For example, TEOS-based silicon dioxide is frequently used in this type of application because the stress of this material is easily controlled at low temperatures (<250°C). The stress of TEOS oxide must be controlled very strictly to achieve a balance between the wafer cracking threshold and warpage.

[0008] Referring to Figures 1(a) and (b) of the drawing, scanning electron microscope (SEM) images of PECVD deposited films with a nominal thickness of 40 μm deposited at 250°C are shown. In both cases, the plasma-facing surface is the top surface of the wafer. In Figure 1(a), delamination is evident in the bevel of the TEOS-based SiO2 film in this case. Figure 1(b) shows ideal deposition using a novel silicon oxycarbide film in which the thickness gradually decreases near the bevel, and a continuous film without delamination exists.

[0009] At the vertical edge of the bevel in Figure 1(a), the TEOS film is thought to be more tensile than the bulk film due to reduced ion bombardment during the deposition process, resulting in thin film delamination. The bevel region is also a source of cracks that can eventually propagate into the bulk of the TEOS oxide film, i.e., beyond the 2 mm edge exclusion region. Since both delamination and cracking are unacceptable, the use of these TEOS oxide films is limited to smaller film thicknesses.

[0010] The introduction of carbon into an SiO matrix is ​​expected to reduce the density and potential stress within the film. However, if excessive carbon is introduced into the film, an increase in the coefficient of thermal expansion is expected. Organic films have a high coefficient of thermal expansion and tend to experience tensile stress, making them impractical for this type of application. The challenge is to introduce sufficient carbon into the SiO matrix while reducing film stress and producing a weakly compressible film with adequate mechanical strength. Therefore, the inventors have devised an improved method that utilizes the use of silicon oxycarbide-containing PECVD films to address the above concerns and extend the crack threshold of these films to films with a thickness of >>40 μm. [Means for solving the problem]

[0011] The present invention provides a method for controlling stress fluctuations in a silicon oxycarbide-containing layer formed using a capacitively coupled parallel plate plasma-enhanced chemical vapor deposition system, the method comprising the following steps: - A step of providing a chamber equipped with a platen for supporting a substrate, - The step of placing a substrate on a platen on which a layer containing silicon oxycarbide can be formed on the substrate, - The step of introducing an oxygen-containing gas into the chamber, - The step of introducing trimethylsilane gas into the chamber, - A step of generating plasma in a chamber, - Maintain the temperature at which deposition occurs below 275°C; - A step of maintaining the pressure inside the chamber in the range of 1.5 to 3 Torr, - A step in which a layer is deposited on a substrate at a deposition rate of >1.5 μm / min.

[0012] TEOS SiO2 generates edge defects when thick films are deposited. These edge defects are undesirable and require additional process steps, as they can often be cleaned. Introducing C into the SiO matrix results in lower-stress films that can achieve very thick films without edge defects. Deposition rates and film uniformity values ​​have been found to be comparable to or better than those of TEOS-based films.

[0013] The oxygen-containing gas may include oxygen gas or nitrous oxide gas. In the first embodiment, the oxygen-containing gas includes oxygen gas, which is introduced into the chamber at a flow rate in the range of 300 to 360 sccm, optionally 333 sccm. In the second embodiment, where the oxygen-containing gas includes nitrous oxide, the nitrous oxide is introduced into the chamber at a flow rate in the range of 1160 to 1560 sccm, optionally 1360 sccm. In the second embodiment, the method may further include the step of introducing argon gas into the chamber, which can be done at a flow rate of argon gas in the range of 254 to 414 sccm, optionally 384 sccm.

[0014] The method further includes introducing trimethylsilane gas into the chamber at a flow rate in the range of 330 to 380 sccm. Optionally, in the first embodiment, trimethylsilane gas is introduced into the chamber at a flow rate of 380 sccm, while in the second embodiment, trimethylsilane gas is introduced into the chamber at a flow rate of 330 sccm.

[0015] This method further includes maintaining the temperature at which deposition occurs within the range of 225–275°C. Optionally, the temperature is maintained at 250°C in relation to both embodiments.

[0016] The plasma can be generated by one or more radio frequency (RF) power supplies that provide mixed-frequency power including high-frequency and low-frequency components. The high-frequency component may include power in the range of 1100 to 1410 W, and the low-frequency component may include power in the range of 0 to 600 W. Optionally, in the first embodiment, the high-frequency component includes power of 1210 W and the low-frequency component includes power of 400 W. Optionally, in the second embodiment, the high-frequency component includes power of 1200 W and the low-frequency component includes power of 0 W.

[0017] In the first embodiment, the layer formed on the substrate contains silicon oxycarbide. In the second embodiment, the layer formed on the substrate contains nitrogen-doped silicon oxycarbide.

[0018] In the first embodiment, the pressure is maintained at substantially 2.5 Torr in the chamber, and in the second embodiment, the pressure is maintained at substantially 1.73 Torr in the chamber.

[0019] The dielectric constant of a silicon oxycarbide film is substantially 3.4. Similarly, the dielectric constant of a nitrogen-doped silicon oxycarbide film is substantially 3.2. [Brief explanation of the drawing]

[0020] [Figure 1]FIG. 1 is a SEM image. (a) is a SEM image of a TEOS oxide-based film showing peeling around the bevel film region, and (b) is a SEM image of a silicon oxycarbide-based film showing no peeling around the bevel film region. [Figure 2] FIG. 2 is a schematic diagram of a system for performing PECVD of a film containing silicon oxycarbide. [Figure 3] FIG. 3 is a flowchart showing steps related to the method according to the present invention. [Figure 4] FIG. 4 is an FTIR graph showing the absorbance spectra of a silicon oxycarbide film (solid line) and a nitrogen-doped silicon oxycarbide film (dotted line) deposited on a silicon wafer, showing signature bonds involved in absorption at various wave numbers.

Embodiments for Carrying Out the Invention

[0021] Referring to Figure 2, a schematic diagram of System 1 for controlling stress fluctuations in a layer (not shown) formed on a substrate by PECVD according to an embodiment of the present invention is shown. System 1 may include an SPTS Delta® fxPECVD system and includes a chamber 2 for electrically grounded processing in which the PECVD process is carried out. Chamber 2 includes a processing area 3 in which a plasma is generated, defined between a substrate support or platen 5 and a gas inlet system including an aluminum showerhead 4. During use, the gas enters the rear side of the showerhead 4 and is radiated into the processing area 3 through an internal gas distribution plate 11 and a series of small holes 10 in the showerhead 4. The platen 5, which is the substrate support, is grounded, and the showerhead 4 is powered by plasma generating means including two different RF power supplies 14, 15 and associated matching units 16. The RF power supplies 14, 15 operate at different RF frequencies, i.e., high frequency and low frequency. The grounded substrate support platen 5 is resistance-heated during use to heat the wafer, and a cooling mechanism (not shown) allows for control of the wafer temperature during the deposition process to maintain the wafer temperature below 275°C.

[0022] A dielectric break (not shown) exists between the aluminum showerhead 4 and the metal chamber 2. In different embodiments, alternative plasma generation means may be used, for example, in a configuration where the platen 5, which is a substrate support, is RF driven and the gas distribution plate 11 is grounded. During use, process gas enters the gas inlet system through pipes 6 and 7 equipped with isolation valves 8 and 9. Multiple gas sources are connected to pipes 6 and 7 by a manifold (not shown) having isolation valves, a mass flow controller and associated components. The grounded chamber 2 is exhausted through an exhaust port 12 by a pump (not shown).

[0023] A typical process cycle begins by loading a substrate 22, such as a silicon wafer having a diameter of 200 mm or 300 mm, into a vacuum-sealed chamber 2 and placing the substrate on a substrate support or platen 5. The substrate is positioned within an annular homogenization ring 13 located on the upper surface of the substrate support. A process gas or gas mixture passes through a gas pipe 6 with the corresponding isolation valve 8 open and enters the chamber 2 through a showerhead 4. Once the desired gas flow and pressure are achieved, a system controller 24 associated with the chamber 2 signals RF power delivered to the showerhead 4 by either or both of the RF power supplies 14, 15, thereby initiating plasma within the chamber 2 between the showerhead 4 and the platen 5, which is the substrate support. As understood in the art, to maintain the pressure in the chamber 2 in the range of 1.5 to 3 Torr, the pressure is controlled by controlling the gas flow in and out of the chamber 2, for example by controlling the pump speed through the exhaust port and / or by using a mass flow controller that controls the flow into the chamber.

[0024] The high-frequency RF power supply 14 operates at 10-20 MHz, and the low-frequency RF power supply 15 operates at 375 kHz-2 MHz. Matching units 16 and 17 enable efficient coupling of RF power to the showerhead 4.

[0025] Once the desired deposition thickness is achieved, the controller 24 emits an off signal, the power is switched off, and the process gas flow is stopped. The controller 24 is coupled to several sensors (not shown) in the chamber 2. The chamber 2 is evacuated as the process gas passes through the exhaust port 12, and then the substrate 22 is removed from the chamber 2.

[0026] High-frequency RF power (13.56 MHz in an exemplary embodiment) is applied across parallel plates in the reactor chamber to generate plasma from the process gas. This is a capacitively coupled plasma-enhanced chemical deposition system having a pair of capacitively coupled parallel plates. The first plate (i.e., electrode) is supplied by a platen 5, which is a grounded wafer support. The second plate (i.e., electrode) is supplied by a powered showerhead 4. While the system 1 can provide mixed-frequency power including high-frequency and low-frequency components, in the exemplary embodiment, the system can be operated at a single frequency only. Relatively low RF power is supplied by a high-frequency RF power supply 15 (at 13.56 MHz in an exemplary embodiment).

[0027] Referring to Figure 3 of the drawings, a flowchart is provided showing steps relating to Method 100 according to one embodiment of the present invention for controlling stress fluctuations in a layer containing silicon oxycarbide formed using PECVD. The method involves the use of System 1 disclosed above and includes a first step 101 of placing a substrate, i.e., a substrate (wafer) 22 on which a silicon oxycarbide layer can be formed, on a platen 5. In the first embodiment, in step 102a, oxygen gas is introduced into the chamber 2 as an oxygen-containing gas, along with trimethylsilane gas, via a gas pipe 6 and a showerhead 4. Then, in step 103a, RF power supplies 14, 15 are activated to generate plasma in the chamber 2 between the showerhead 4 and the platen 5 in order to start the PECVD process. In the second embodiment, in step 102b, the oxygen gas is replaced with nitrous oxide gas, and the oxygen-containing gas containing nitrous oxide gas is also introduced into the chamber 2 along with trimethylsilane gas, via a gas pipe 6 and a showerhead 4. Next, in step 103b, the RF power supplies 14 and 15 are activated in the same manner to generate plasma in the chamber 2 between the showerhead 4 and the platen 5 in order to initiate the PECVD process.

[0028] In the first embodiment, a silicon oxycarbide film is deposited on the substrate (wafer) 22, whereas in the second embodiment, a nitrogen-doped silicon oxycarbide film is deposited on the substrate (wafer) 22. In both embodiments, the film is deposited at a deposition rate of >1.5 μm / min, resulting in a film thickness in the range of 10 to 90 μm. Once the desired thickness is achieved, in step 104, the process gas flow into the chamber 2 is stopped, and the gas in the chamber is exhausted through the exhaust port 12. Next, in step 105, the substrate 22 is removed from the chamber 2.

[0029] The process conditions related to the first and second embodiments are shown in Table 1 below: [Table 1]

[0030] Table 2 below outlines the characteristics of wafers having the deposited films according to the first and second embodiments, along with the TEOS oxide film for comparison:

[0031] [Table 2]

[0032] Table 2 clearly shows that the SiCON film provides the lowest non-uniformity value, and acceptable uniformity is achieved in all cases. Further improvement in uniformity values ​​is expected to be achievable without adversely affecting other film properties.

[0033] Furthermore, it can be observed that the refractive index gradually increases from the TEOS film to the SiCON and SiCO films. Most notably, with respect to thicker film deposition, a decrease in film stress is observed from TEOS to SiCO, with the lowest compressive stress being found in the SiCON film. These stress measurements were performed on 1 μm films. The stress values ​​for crack-free 40 μm SiCO and SiCON films were -24 MPa and -1 MPa, respectively. The elastic modulus and hardness of the films are lower in the SiCO and SiCON films compared to the TEOS film. In addition, both film deformations clearly demonstrated excellent wafer bevel defect properties when thick films (40 μm) were deposited. Improvement in wafer bevel defects is desirable for positioning devices closer to the wafer edge.

[0034] FTIR analysis of both films was also performed using a Nicolet iS50 FT-IR spectrometer, and Figure 4 provides a graph showing the absorbance spectra of both the SiCO and SiCON films. The SiCO FTIR spectrum is identified by a solid line, and the SiCON FTIR spectrum by a dotted line. Unlike the TEOS film, which is dominated by a Si-O peak at approximately 1025 cm⁻¹, the SiCO film also has a Si-CHx peak at 2965–2910 cm⁻¹, a Si-C peak at 1271 cm⁻¹, and a Si-Hx peak at approximately 2280–2170 cm⁻¹. The SiCON film also has a weak NH peak at 3390 cm⁻¹ and a SiN peak at 840 cm⁻¹, indicating evidence of N incorporation.

[0035] The Si-CHx and SiC peaks are more pronounced in the case of SiCON than in the case of SiC films, suggesting that the carbon content of these films is higher than that of equivalent SiCO films. The films are silicon oxycarbide hydrogenates with a trimethylsilane / N2O film containing N.

[0036] Surprisingly, the dielectric constant k-value of the optimized SiCO film was 3.4, in contrast to the k-value of 5.8 achieved by Laboda et al. (US6593655), which suggests an increase in carbon incorporation into the film. The dielectric constant k-value of the SiCON film was 3.2, suggesting a slightly higher carbon content compared to the SiCO film, as indicated by the FTIR Si-CHx and SiC peaks.

[0037] SiCO films with a maximum thickness of 40 μm and SiCON films with a maximum thickness of 90 μm showed no cracking or delamination around the bevel.

[0038] From the above, it is clear that the method disclosed herein provides an improved approach to PECVD deposition of silicon oxycarbide-containing films. [Explanation of Symbols]

[0039] 1 System, 2 Chamber, 3 Processing area, 4 Shower head, 5 Platen, 6,7 Pipes, 8,9 Isolation valves, 10 Hole, 11 Gas distribution plate, 12 Exhaust port, 13 Annular homogenizing ring, 14,15 Power supply.

Claims

1. A method for controlling stress fluctuations in a silicon oxycarbide-containing layer formed using a capacitively coupled plasma-enhanced chemical deposition system, This delicious, - A step of providing a chamber equipped with a platen for supporting a substrate, - A step of placing a substrate on a platen that can form a layer containing silicon oxycarbide on the substrate, - The step of introducing an oxygen-containing gas into the chamber, - The step of introducing trimethylsilane gas into the chamber, - A step of generating plasma in the chamber, - A step of maintaining the temperature at which deposition occurs below 275°C, - A step of maintaining the pressure inside the chamber in the range of 1.5 to 3.5 Torr, - A step of depositing a layer on a substrate at a deposition rate > 1.5 μm / min, A method that includes this.

2. The method according to claim 1, wherein the oxygen-containing gas includes oxygen gas.

3. The method according to claim 2, wherein oxygen gas is introduced into the chamber at a flow rate in the range of 300 to 360 sccm.

4. The method according to claim 1, wherein the oxygen-containing gas includes nitrous oxide.

5. The method according to claim 4, wherein nitrous oxide gas is introduced into the chamber at a flow rate in the range of 1160 to 1560 sccm.

6. The method according to claim 4 or 5, further comprising introducing argon gas into the chamber.

7. The method according to claim 6, wherein argon gas is introduced into the chamber at a flow rate in the range of 354 to 414 sccm.

8. The method according to any one of claims 1 to 7, wherein trimethylsilane gas is introduced into the chamber at a flow rate in the range of 330 to 380 sccm.

9. The method according to any one of claims 1 to 8, wherein the deposited layer has a thickness in the range of 10 to 90 μm.

10. The method according to any one of claims 1 to 9, wherein the temperature inside the chamber is maintained in the range of 225 to 275°C.

11. The method according to any one of claims 1 to 10, wherein the temperature inside the chamber is maintained at substantially 250°C.

12. The method according to any one of claims 1 to 11, wherein the plasma is generated using one or more RF power supplies having mixed frequency power including high-frequency and low-frequency components.

13. The method according to claim 11, wherein the high-frequency component includes power in the range of 1100 to 1410 W.

14. The method according to claim 12 or 13, wherein the low-frequency component includes power in the range of 0 to 600 W.

15. The method according to any one of claims 1 to 14, wherein the layer comprises silicon oxycarbide.

16. The method according to any one of claims 1 to 14, wherein the layer comprises nitrogen-doped silicon oxycarbide.

17. The method according to claim 15, wherein the dielectric constant of the silicon oxycarbide film is substantially 3.

4.

18. The method according to claim 16, wherein the dielectric constant of the nitrogen-doped silicon oxycarbide film is substantially 3.2.

Images