Impact of H-Related Chemical Bonds on Physical Properties of SiN_x:H Films Deposited via Plasma-Enhanced Chemical Vapor Deposition

Abstract

SiN_x:H film deposition via plasma-enhanced chemical vapor deposition has been widely used in semiconductor devices. However, the relationship between the chemical bonds and the physical and chemical properties has rarely been studied for films deposited using tools in terms of the actual volume production. In this study, we investigated the effects of the deposition conditions on the H-related chemical bonding, physical and chemical properties, yield, and quality of SiN_x:H films used as passivation layers at the 28 nm technology node. The radiofrequency (RF) power, electrode plate spacing, temperature, chamber pressure, and SiH₄:NH₃ gas flow ratio were selected as the deposition parameters. The results show a clear relationship between the H-related chemical bonds and the examined film properties. The difference in the refractive index (RI) and breakdown field (E_B) of the SiN_x:H films is mainly attributed to the change in the Si–H:N–H ratio. As the Si–H:N–H ratio increased, the RI and E_B showed linear growth and exponential downward trends, respectively. In addition, compared with the Si–H:N–H ratio, the total Si–H and N–H contents had a greater impact on the wet etching rates of the SiN_x:H films, but the stress was not entirely dependent on the total Si–H and N–H contents. Notably, excessive electrode plate spacing can lead to a significant undesired increase in the non-uniformity and surface roughness of SiN_x:H films. This study provides industry-level processing guidance for the development of advanced silicon nitride film deposition technology.

1. Introduction

As the core of information technology development, the integrated circuit (IC) industry involves hundreds of processes, including lithography, etching, film deposition, etc. As one of the three core processes, the film deposition process includes physical vapor deposition, chemical vapor deposition, epitaxy, etc. [1]. Plasma-enhanced chemical vapor deposition (PECVD) is a form of chemical vapor deposition and has been widely used in many fields. According to the various classifications, PECVD encompasses a wide range of species. Based on the power supply utilized for the plasma excitation, PECVD can be categorized into radiofrequency PECVD (RF-PECVD), very high-frequency PECVD (VHF-PECVD), electron cyclotron resonance PECVD (ECR-PECVD), and linear microwave PECVD (LM-PECVD) [2,3]. Depending on the sample placement structure, either tube PECVD or plate PECVD is employed. Due to the excellent electrical properties, substrate adhesion, and step coverage of the deposited film, the utilization of RF power and flat-panel structures is prevalent in PECVD devices in ultra-large-scale IC fields [4,5]. Among the various films, silicon nitride films deposited via PECVD have been widely used in semiconductor devices including logic and memory devices because of their high dielectric constant, stable chemical properties, and strong resistance to sodium ion and water vapor [6]. For example, in the production and preparation process of the memory/storage circuit, silicon nitride films can be used as stress lining layers to enhance the carrier mobility and improve the electrical performance. Moreover, as passivation layers, silicon nitride films can prevent other layers from being eroded by the etching solution, and they can also be used as hard masks, which can cooperate with the photoresist to form mask patterns [7,8,9,10]. Therefore, silicon nitride films deposited via PECVD have attracted the attention of many researchers [11,12].

Currently, empirical research has indicated that various deposition parameters, such as radiofrequency (RF) power, chamber pressure, and temperature, have distinct effects on the properties of silicon nitride film synthesized via PECVD [12,13]. For instance, when the SiH₄:NH₃ gas flow ratio decreases, the breakdown electric field is improved [14]. The elastic modulus and hardness increase with a decrease in the chamber pressure [15]. Moreover, as the RF power decreases, the refractive index (RI) increases [16]. The influence mechanisms of some deposition conditions on the film performance have also been studied [17,18]. The hydrogen concentration is decreased and the density is increased by increasing the low-frequency power, thereby improving the mobility, on–off ratio, leakage current, and subthreshold of the film [19]. The N:Si ratio decreases with an increase in the SiH₄ flow rate, which leads to a decrease in the RI [20]. The diffusion rate of the reactants is improved by raising the substrate temperature, resulting in denser films [21]. These studies provide some theoretical guidance for the production and application of silicon nitride films; however, many were conducted using laboratory-level instruments rather than the mature tools used in actual volume production. Therefore, the process parameters selected in these studies should be considered for reference only. There is a considerable gap between the laboratory research conclusions and the experiences gained in actual production and application. Moreover, at the industry level, there is a serious lack of deep insight into the effect mechanisms of the different deposition conditions on the film properties. To meet the standards for silicon nitride film properties, a series of orthogonal and univariate tests are blindly performed, resulting in a significant waste of time and money [22].

In this study, to systematically assess the relationship between the deposition conditions, H-related chemical bonds, and properties, silicon nitride films were deposited via mature machines under the corresponding deposition conditions and were used as the passivation layers for a 28 nm technological node. Subsequently, the effect mechanisms of the different deposition conditions on the H-related chemical bonds and film properties were explored, including the RF power, electrode plate spacing, temperature, chamber pressure, and SiH₄:NH₃ gas flow ratio. In addition, the relationships between the H-related chemical bonds and film properties, including the RI, breakdown field strength, stress, and wet etching rate, were studied. In addition, considering the yield and quality, the effects of the different deposition conditions on the deposition rate (DR), non-uniformity (NU), and surface roughness of the films were also investigated. The results provide potential theoretical guidance for the development of advanced silicon nitride film deposition technology.

2. Film Deposition and Testing Methods

2.1. Film Deposition

The SiN_x:H films were deposited onto the substrate surfaces via PECVD equipment (Piotech, Ltd., Shenyang, China). The substrates were 12 in. p-type (100) silicon wafers. The reaction gases included SiH₄, NH₃, and N₂, and the clean gases included NF₃, He, and N₂. The purities of these gases were greater than 99.99%. A schematic diagram of the SiN_x:H film deposition process is shown in Figure 1 and is as follows [23]: (a) the SiH₄, NH₃, and N₂ gases are transmitted to the chamber in a certain proportion; (b) the gas molecules are then converted into plasmas via an RF electric field; (c) these plasmas collide and react with each other to produce precursors; (d) some of the precursors adhere to the silicon wafer surface and diffuse; (e) these precursors react and nucleate on the silicon wafer surface; and (f) the SiN_x:H films are formed through a series of chemical reactions.

The specific chemical reactions can be described as follows [24]: Under the influence of a radiofrequency electric field, NH₃ and N₂ undergo electrolysis to produce N, NH, and NH₂, while SiH₄ is electrolyzed to yield SiH_n. These reactive groups react with each other to form precursors, including SiH_m(NH_n), Si(NH₂)₄, Si(NH₂)₃, etc. Under an electric field of low radiofrequency power, ethylsilane is the main product. However, under an electric field of high radiofrequency power, Si(NH₂)₄ and Si(NH₂)₃ are the main products. These amino silanes are the main precursors of silicon nitride films and are closely related to the film properties. Changes in the deposition conditions will affect the chemical process, thereby affecting the film properties.

Silicon nitride film deposited via PECVD includes some Si–H and N–H, and the Si:N atomic ratio is not a standard chemical ratio (Si:N = 3:4). Therefore, the hydrogenated silicon nitride component is denoted as Si_xN_yH_z, often simply referred to as SiN_x:H film. The chemical reaction equation is shown in Equation (1):

$$\mathrm{S}\mathrm{i}{\mathrm{H}}_4+\mathrm{N}{\mathrm{H}}_3\to \mathrm{S}{\mathrm{i}}_{\mathrm{x}}{\mathrm{N}}_{\mathrm{y}}{\mathrm{H}}_{\mathrm{z}}$$

(1)

To explore the effect mechanisms of these deposition conditions on the H-related chemical bonds and SiN_x:H film properties, a series of univariate experiments are carried out. According to our team’s previous study and the relevant literature [6,25], the basic deposition conditions are as shown in

Table 1. Basic deposition conditions for SiN_x:H films.

Deposition Conditions	RF Power	Electrode Plate Spacing	Temperature	Pressure	SiH4:NH3 Gas Flow Ratio
	(W)	(mm)	(°C)	(Torr)
Value	1230	19.0	400	3.30	3.20

. Under these deposition conditions, the deposited SiN_x:H films can be used as the passivation layers for a 28 nm technological node. On this basis, as shown in

Table 2. Experimental conditions for SiN_x:H film deposition.

Deposition Conditions	RF Power	Electrode Plate Spacing	Temperature	Pressure	SiH4:NH3 Gas Flow Ratio
	(W)	(mm)	(°C)	(Torr)
Basic	1230	19.0	400	3.30	3.20
RF Power	1082	19.0	400	3.30	3.20
	1180	19.0	400	3.30	3.20
	1144	19.0	400	3.30	3.20
	1107	19.0	400	3.30	3.20
	1279	19.0	400	3.30	3.20
	1300	19.0	400	3.30	3.20
Electrode Plate Spacing	1230	17.1	400	3.30	3.20
	1230	17.7	400	3.30	3.20
	1230	18.2	400	3.30	3.20
	1230	19.8	400	3.30	3.20
	1230	20.3	400	3.30	3.20
	1230	20.9	400	3.30	3.20
Temperature	1230	19.0	360	3.30	3.20
	1230	19.0	372	3.30	3.20
	1230	19.0	384	3.30	3.20
	1230	19.0	400	3.30	3.20
	1230	19.0	416	3.30	3.20
	1230	19.0	428	3.30	3.20
	1230	19.0	440	3.30	3.20
Pressure	1230	19.0	400	3.07	3.20
	1230	19.0	400	3.17	3.20
SiH4:NH3 Gas Flow Ratio	1230	19.0	400	3.30	2.98
	1230	19.0	400	3.30	3.07
	1230	19.0	400	3.30	3.33
	1230	19.0	400	3.30	3.42
	1230	19.0	400	3.30	3.51

, the ranges of the changes in the RF power, electrode plate spacing, temperature, chamber pressure, and SiH₄:NH₃ gas flow ratio are from 1082 to 1300 W, from 17.1 to 20.9 mm, from 360 to 440 °C, from 3.07 to 3.30 Torr, and from 2.98 to 3.51, respectively.

2.2. Characterization Methods

2.2.1. H-Related Chemical Bonds

The H-related chemical bonds of the SiN_x:H films were analyzed via Fourier-transform infrared spectroscopy (FTIR) (Nicolet, Massachusetts, USA). The spectra were acquired in the wavenumber range of 4000 to 400 cm⁻¹ with a resolution of 2 cm⁻¹. Based on the test spectrum, the Si–H or N–H bond density in the film (ρ_Si(N)–H) was calculated using Equation (2). Among all the atoms, the atomic percentage of H in Si–H and N–H (H_Si(N)–H) was calculated using Equation (3), the total atomic percentage of H in Si–H and N–H (H_Total) was calculated using Equation (4), and the Si–H and N–H bond density ratio (R_Si/N) was calculated using Equation (5) [26]:

$${\rho }_{Si(N)–H}({\mathrm{atom}/\mathrm{cm}}^{-3})=\frac{S_{S\mathrm{i}(N)–H}\times {10}^8}{{\sigma }_{Si(N)–H}\times THK}$$

(2)

$$H_{Si(N)–H}(\%)=\frac{{\rho }_{S\mathrm{i}\left(N\right)–H}}{6.0108\times {10}^{22}+{\rho }_{S\mathrm{i}–H}+{\rho }_{N–H}}\times 100\%$$

(3)

$$H_{T\mathrm{o}tal}(\%)=\frac{{\rho }_{S\mathrm{i}–H}+{\rho }_{N–H}}{6.0108\times {10}^{22}+{\rho }_{S\mathrm{i}–H}+{\rho }_{N–H}}\times 100\%$$

(4)

$$R_{S\mathrm{i}/N}=\frac{{\rho }_{S\mathrm{i}–H}}{{\rho }_{\mathrm{N}–H}}$$

(5)

where S_Si(N)–H is the absorption peak area of the Si–H or N–H in the film, σ_Si(N)–H is the Si–H or N–H bond cross section, σ_Si–H and σ_N–H are 7.4 × 10⁻¹⁸ and 5.3 × 10⁻¹⁸ cm², and THK is the film thickness.

2.2.2. Refractive Index

The RI of the SiN_x:H film was measured via the film metrology system with a wavenumber range of 200 to 800 cm⁻¹ (Tensor 8500, KLA, Milpitas, CA, USA). The RI data were collected from five wafers, and 49 points were measured for each wafer.

2.2.3. Breakdown Field Strength

The breakdown field strength of the film was measured via the mercury CV test method within a field strength range of -150 to 50 V cm⁻¹ (MC530, Semilab, Hungary). The breakdown field strength (E_B) was collected from five wafers, with five points measured for each wafer, and it was calculated using Equation (6) and determined by testing five films [27]:

$$E_B(\mathrm{MV}/\mathrm{cm})=E_1+\frac{\left(J_B-J_1\right)\times (E_2-E_1)}{J_2-J_1}$$

(6)

where J_B is 10⁻³ A/cm² (when the current densities are lower than J_B, the films have broken down), E₂ and E₁ are the electric field strengths of the test points before and after the breakdown, respectively, and J₂ and J₁ are the current densities of the test points before and after the breakdown, respectively.

2.2.4. Stress

The curvature radius of the wafer was measured via T910 (Skyverse Technology, China). The test mode was a line scan with 51 points. Then, the internal stress (σ_f) of the film was calculated using the Stoney equation (Equation (7)):

$${\sigma }_f(\mathrm{MPa})=\frac{E_s{t}_s}{6(1-V_s)THK}(\frac{1}{R_f}-\frac{1}{R_s})$$

(7)

where E_s, t_s, and V_s are Young’s modulus, Poisson’s ratio, and thickness of the silicon wafer, respectively, and R_s and R_f are the curvature radii of the wafer before and after the film deposition. The stress data were collected from the measurements of five wafers.

2.2.5. Wet Etching Rate Ratio

Hydrofluoric acid with a concentration of 1 wt.% was utilized for the wet etching rate tests, realized by diluting 49 wt.% hydrofluoric acid (Aladdin, MW = 20.01 g/mol) in DI water. The wet etching rate reflected the rate of the reaction of the film with the hydrofluoric acid solution. The experiment was performed at room temperature utilizing a Teflon beaker as the container. However, concentration differences in the hydrofluoric acid solutions prepared at different times were observed. Therefore, to more accurately compare the wet etching rates of the different films, the deposited films were etched together with the reference samples. The wet etching rate ratio (WERR) was calculated using Equation (8) and was collected from the measurements of five wafers:

$$WERR=\frac{TH{K}_{f-pre}-TH{K}_{f-post}}{TH{K}_{r-pre}-TH{K}_{r-post}}$$

(8)

where THK_f-pre is the original thickness of the film, THK_f-post is the thickness of the film after the wet etching, THK_r-pre is the original thickness of the reference samples, and THK_r-post is the thickness of the reference samples after the wet etching.

2.2.6. Deposition Rate and Non-Uniformity

The calculation of the DR and NU of the SiN_x:H films mainly depended on the THK, which was also measured via the KLA Tensor 8500 with a wavenumber range of 200 to 800 cm⁻¹. The DR and NU were calculated using Equations (9) and (10), respectively, and were collected from five wafers, with 49 points measured for each wafer:

$$DR(\AA /\mathrm{s})=\frac{THK}{t_{Dep}}$$

(9)

$$NU(\%)=\frac{\sqrt{{\displaystyle \sum _1^{49}{(TH{K}_{\mathrm{Point}-n}-THK)}^2}}}{THK}$$

(10)

where THK is the mean of the thicknesses of 49 points, THK_Point-n is the thickness of the nth point, and t_Dep is the deposition time.

2.2.7. Surface Roughness

The surface roughness of the film was tested with an atomic force microscope (Agilent, Santa Clara, CA, USA). The sample needed to be processed into a size of 2 × 2 cm². The RMS was used to evaluate the surface roughness of the film, which is the square root of the square sum of the average contour deviations. The DR and NU were collected from three wafers.

3. Results and Discussion

3.1. Influence of Deposition Conditions on H-Related Chemical Bonds

For SiN_x:H films, the RI, WERR, stress, and other properties are closely related to the H-related chemical bonds (Si–H and N–H), which are dependent on the deposition conditions [28]. The SiN_x:H films deposited under the different conditions exhibited varying H_Si–H, H_N–H, R_Si/N, and H_Total values. To understand how the deposition conditions impacted the film properties by influencing the H-related chemical bonds, in this section, we investigate the effects of the deposition conditions, including the RF power, electrode plate spacing, temperature, chamber pressure, and SiH₄:NH₃ gas flow ratio, on H_Si–H, H_N–H, R_Si/N, and H_Total.

3.1.1. RF Power

Reactive radicals, ions, neutral atoms, molecules, etc., from reaction gas can be converted by RF power, affecting the atom bonding structures [29]. The FTIR spectra of films deposited under different RF powers are shown in Figure 2a. The absorption peak at ca. 2183 cm⁻¹ is due to the Si–H stretching vibration, while the peak at ca. 3345 cm⁻¹ is ascribed to N–H stretching [30]. According to the FTIR spectra, the H_Si–H, H_N–H, R_Si/N, and H_Total values are as shown in Figure 2b,c. SiH₄ and NH₃ can be more fully dissociated by increasing the RF power. SiH₄ can be easily dissociated into SiH₃⁺, SiH₂⁺, and SiH⁺; thus, the increase in the RF power mainly promotes the reaction of SiH₃⁺ and SiH₂⁺ dissociation into SiH⁺. For NH₃, the N–H bond energy is higher than the Si–H bond energy (390.8 vs. 318.0 kJ/mol); thus, NH₃ dissociation is more difficult than SiH₄ dissociation, and the increase in the RF power mainly promotes the dissociation of NH₃ [31]. Therefore, as the RF power increases, H_Si–H decreases while H_N–H increases (Figure 2b), and the decrease in H_Si–H is greater than the increase in H_N–H. Therefore, as the RF power increases, R_Si/N and H_Total decrease.

3.1.2. Electrode Plate Spacing

The effect of electrode plate spacing on the FTIR spectra, and the H_Si–H, H_N–H, R_Si/N, and H_Total values, of the films is shown in Figure 3. As the electrode plate spacing increases, the time it takes for SiH₄ and NH₃ to migrate from the top electrode to the silicon wafer increases, which can lead to a longer SiH₄ and NH₃ residence time in the electric field and, subsequently, to a fuller dissociation. Thus, like the RF power, the increase in the electrode plate spacing mainly promotes the reaction of SiH₃⁺ and SiH₂⁺ dissociating into SiH⁺ and the dissociation of NH₃. Therefore, as can be seen in Figure 3b,c, as the electrode plate spacing increases, H_Si–H decreases while H_N–H increases. Meanwhile, the decrease in H_Si–H is also greater than that in H_N–H. Therefore, as the electrode plate spacing increases, R_Si/N and H_Total decrease.

3.1.3. Temperature

Figure 4a shows the FTIR spectra of films under different deposition temperatures. The evolution trends of H_Si–H, H_N–H, R_Si/N, and H_Total with the temperature are plotted in Figure 4b,c. With the increase in the deposition temperature, H_Si–H, H_N–H, and H_Total decreased, which indicates that Si–H and N–H breakdown can be accelerated by increasing the deposition temperature [32]. Although the N–H bond energy was higher than the Si–H bond energy, as the temperature increased, the downward trend for H_Si–H was significantly lower than that for H_N–H, and R_Si/N increased (Figure 4b,c). The reason for this is that more SiH₄ was involved in the film deposition process due to the increased temperature, which made it easier for the Si atoms to bond.

3.1.4. Chamber Pressure

Figure 5a presents the FTIR spectra of films deposited under different pressures. Through the peak areas, the evolution trends of H_Si–H, H_N–H, R_Si/N, and H_Total with the pressure were calculated and are shown in Figure 5b,c. As a result of the chamber pressure increase, the mean free paths of the gases were reduced. This reduction can lead to a decrease in the distance over which the acceleration of molecules, ions, electrons, and other particles occurs [33]. Therefore, as the chamber pressure increases, the energy and activity of these particles are reduced, and the degree of SiH₄ and NH₃ dissociation decreases. Meanwhile, due to the decrease in the mean free paths of the gas, the bonding of the Si atoms is stronger. Therefore, as the chamber pressure increases, H_Si–H decreases, and H_N–H and R_Si/N increase. The decrease in H_Si–H is also greater than the increase in H_N–H; thus, H_Total decreases with the increase in the chamber pressure.

3.1.5. SiH₄:NH₃ Gas Flow Ratio

We evaluated the impact of the SiH₄:NH₃ gas flow ratio on the H-related chemical bonds via FTIR, and the results are shown in Figure 6a. Different gas flow ratios were obtained by keeping the NH₃ flow constant and adjusting the SiH₄ flow. Figure 6b,c show that as the SiH₄:NH₃ gas flow ratio increased, H_Si–H and R_Si/N increased, while H_N–H decreased. These evolution trends are consistent with previously reported results [31]. Per unit volume, the SiH₄ count increased and, accordingly, the NH₃ count decreased. Meanwhile, the increase in H_Si–H was also greater than the decrease in H_N–H; thus, H_Total increased with the increase in the SiH₄:NH₃ gas flow ratio.

3.2. Influence of Deposition Conditions on SiN_x:H Film Properties

The SiN_x:H films deposited under different conditions have different H_Si–H, H_N–H, R_Si/N, and H_Total values. The effect mechanisms of H-related chemical bonds on the SiN_x:H film properties, including the RI, E_B, stress, and WERR, are explored in this section, and some quantitative relationships between the H-related chemical bonds and properties are established. When facing actual production needs, these findings can be used to guide the adjustment directions of the deposition conditions, which can substantially reduce the time and economic costs.

3.2.1. Refractive Index

The RI is an important parameter of the optical properties of films, and it can directly affect the lithography process. The effects of different deposition conditions, including the RF power, electrode plate spacing, temperature, chamber pressure, and gas flow ratio, on the RI of SiN_x:H films were explored. As shown in Figure 7a, with an increase in the RF power from 1050 to 1300 W, the RI decreased from 2.100 to 2.028, and with an increase in the electrode plate spacing from 17.1 to 20.9 mm, it decreased from 2.084 to 2.002. Meanwhile, as the RF power or electrode plate spacing increased, R_Si/N decreased. In addition, as can be seen in Figure 7b, the RI increased with an increase in temperature. With an increase in temperature from 360 to 440 °C, the RI increased from 2.009 to 2.071. Meanwhile, due to the increase in temperature, the Si atoms bonded more easily to each other, which led to an increase in R_Si/N. On the contrary, as the chamber pressure increased, R_Si/N decreased. With an increase in the chamber pressure from 3.07 to 3.30 Torr, the RI decreased from 2.065 to 2.044. As can be seen in Figure 7c, as the gas flow ratio increased, R_Si/N and the RI increased. With the increase in the gas flow ratio from 2.98 to 3.51, the RI increased from 2.019 to 2.074.

As mentioned above, with the changes in the deposition conditions, the trends for RI and R_Si/N were always consistent. To further clarify the relationship between the RI and R_Si/N, RI and R_Si/N scatter plots are shown in Figure 7d. The fitted linear formula was used to quantitatively describe the relationship between the RI and R_Si/N. The correlation coefficient (R²) of this formula was 0.91734, confirming that the RI is related to R_Si/N, which indicates that the difference in the RI is mainly due to the change in R_Si/N under different deposition conditions. The reason for this is that the relationship between R_Si/N and the Si:N atomic ratio is positive and linear, and the Si:N atomic ratio and RI are closely related. When the Si:N atomic ratio increases, the Si–Si bonding of the film increases, and the optical properties become closer to those of amorphous silicon [34,35]. Therefore, the RI increases with an increase in R_Si/N. Moreover, the effect of temperature on the RI is more pronounced compared that of to the other deposition conditions, which can be related to H_total. The networked structure of the film can be interrupted by Si–H and N–H. The lower the H_total value, the denser the film structure, and the greater the RI. The effect of temperature on H_total is the greatest among these deposition conditions. As the temperature increases, H_total decreases and R_Si/N increases. Therefore, the RI is the most affected by temperature.

3.2.2. Breakdown Field Strength

The E_B value is important for the working performance of SiN_x:H film, and it is closely related to the reliability of the IC. The effects of different deposition conditions on the E_B of SiN_x:H films were explored. As shown in Figure 8a, with an increase in the RF from 1050 to 1300 W, E_B increased from 3.75 to 6.33, and with an increase in the electrode plate spacing from 17.1 to 20.9 mm, it increased from 4.14 to 7.45 MV/cm. Correspondingly, as the RF power or electrode plate spacing increased, R_Si/N decreased. Moreover, as can be seen in Figure 8b, E_B decreased with an increase in temperature and increased with an increase in chamber pressure. With an increase in temperature from 360 to 440 °C, the breakdown voltage decreased from 6.43 to 4.19 MV/cm, and with an increase in chamber pressure from 3.07 to 3.30 Torr, the breakdown voltage increased from 4.74 to 5.10 MV/cm. Moreover, as the temperature increased, R_Si/N increased, and as the chamber pressure increased, R_Si/N decreased. In addition, with an increase in the gas flow ratio, E_B increased and R_Si/N decreased. As shown in Figure 8c, with an increase in the gas flow ratio from 2.98 to 3.51, the breakdown voltage decreased from 5.74 to 4.20 MV/cm.

Unlike the RI, with the changes in the deposition conditions, the trends for E_B and R_Si/N were always opposite. As R_Si/N increased, E_B decreased. To further explore their relationship, E_B and R_Si/N scatter plots are shown in Figure 8d. The fitted index formula was used to quantitatively describe the relationship between E_B and R_Si/N. The correlation coefficient (R²) of this formula was 0.85614, confirming that the difference in E_B was also mainly due to the change in R_Si/N under different deposition conditions. The reason for this is that as R_Si/N decreases, the bandgap width increases, increasing E_B [36,37]. In addition, temperature has a more obvious impact on E_B, which is also related to H_total [38]. The suspension bonds and other defects in the film can lead to a decrease in E_B, which can be reduced via the introduction of hydrogen [39]. When the deposition temperature increases, H_total decreases, and the trap density and gap state density in the film increase.

3.2.3. Stress

The stress of the film, which indicates the degree of wafer warping, is primarily divided into intrinsic stress and thermal stress. Intrinsic stress is generated during the deposition process and may be introduced due to lattice mismatch, crystal defects, differences in the unit cell parameters, etc., making it difficult to quantify [40]. Meanwhile, the films deposited via PECVD are all SiN_x:H films; thus, the difference in the intrinsic stresses of the films under different deposition conditions is limited. Therefore, the changes in film stress under varying deposition conditions are mainly due to thermal stress, which is influenced by the temperature and can be quantified. The thermal stress (σ_T) can be calculated using Equation (11) [41]:

$${\sigma }_T=\frac{E_f}{1-V_f}{\displaystyle {\int }_{T_{Dep}}^{T_R}({\alpha }_S-{\alpha }_f)}dT$$

(11)

where E_f and V_f are the Young’s modulus and Poisson’s ratio of the film, respectively; T_Dep and T_R are the deposition temperature and test temperature of the film, respectively; α_s and α_f are the thermal expansion coefficients of the silicon wafers and films, respectively.

As can be seen from Equation (11), σ_T, T_Dep and α_f are the critical parameters. In this section, the effects of different deposition conditions on the stresses of SiN_x:H films are discussed. As shown in Figure 9, the stresses of all the films were negative and were compressive stress. The reason for this is that α_f was lower than α_s, and the shrinkage of the silicon wafer was greater than that of the film during the cooling process [40].

Under different RF powers, electrode plate spacings, chamber pressures, and gas flow ratios, T_Dep was the same; thus, the change in the stress can be mainly attributed to the change in α_f. α_f is related to the degree of structure compactness: The denser the film structure, the lower the α_f value and the stress. As can be seen in Figure 9a, as the RF power increased, the stress gradually decreased. With an increase in RF power from 1082 to 1300 W, the stress decreased from −45 to −333 MPa. The reason for this is that as the RF power increases, the molecules are more thoroughly dissociated, and the migration capacity of the active groups increases.

As a result, the structure of the film becomes denser with reduced α_f and stress. Unlike the RF power, as the electrode plate spacing increases, so does the stress of the film. With an increase in electrode plate spacing from 17.1 to 20.9 mm, the stress increased from −308 to −156 MPa. The reason for this may be that the distance from the wafer surface to the pumping pipeline was increased by the increased electrode plate spacing, which caused a delay in the pumping of the byproducts away from the wafer surface. An increase in the concentration of byproducts on the wafer surface leads to a decrease in the concentration of reactants, which can cause the structure of the film to become loose.

As shown in Figure 9b, the stress increased with an increase in chamber pressure. With an increase in chamber pressure from 3.07 to 3.30 Torr, the stress increased from −280 to −214 MPa. The reason for this is that the mean free path of the reactants was reduced with increasing pressure, and the diffusion of the reactants was restricted. Therefore, the film structure became loose and α_f increased. Additionally, as can also be seen in Figure 9c, as the gas flow ratio increased, the stress also increased. With an increase in the gas flow ratio from 298 to 3.51, the stress increased from −305 to −201 MPa, which indicates that α_f increases with the increase in the gas flow ratio.

Under different deposition temperatures, the film stress was related to T_Dep and α_f. The migration capacity of the active groups can be increased by increasing the temperature, which can cause the film structure to become dense and α_f to decrease. The film stress increased with an increase in T_Dep but decreased with a decrease in α_f. As shown in Figure 9b, the stress decreased with an increase in temperature, which indicates that the decrease in α_f played a predominant role. With an increase in temperature from 360 to 440 °C, the stress decreased from −156 to −342 MPa.

Some previous research results showed that the higher the H_Total value, the looser the film structure and the greater the stress [42,43]. To verify this conclusion, we also explored the relationship between H_Total and stress in this research. However, unlike in previous studies, the stress did not increase with elevated H_Total (Figure 9d), which indicates that the loose films were caused not by the increase in H_Total but by other factors, such as the decrease in the reactant diffusion capacity.

3.2.4. Wet Etching Rate Ratio

We explored the effects of different deposition conditions on the WERR of SiN:H film, which is related to R_Si/N and H_Total [44]. As can be seen in Figure 10a, the WERR does not show a constant upward or downward trend with an increase in RF power or electrode plate spacing, which is because R_Si/N and H_Total decrease with an increase in the RF power and electrode plate spacing. As R_Si/N decreases, the WERR increases. However, H_Total decreases with the decreasing WERR [45]. Therefore, the WERR tends to fluctuate with increases in the RF power and electrode plate spacing.

R_Si/N increased while H_Total decreased with the temperature. Therefore, as shown in Figure 10b, the WERR shows a downward trend. With an increase in temperature from 360 to 440 °C, the WERR decreased from 0.66 to 0.56. Unlike the temperature, the WERR shows a fluctuating trend with an increase in chamber pressure. The reason for this is that R_Si/N and H_Total decreased with the increase in chamber pressure.

In addition, as can be seen in Figure 10c, as the gas flow ratio increased, both R_Si/N and H_Total increased, and the WERR shows an upward rather than fluctuating trend. This indicates that, compared with the increase in R_Si/N, the increase in H_Total played the dominant role. With an increase in the gas flow ratio from 2.98 to 3.51, the WERR increased from 0.57 to 0.62. The relationships between H_Total and WERR and between R_Si/N and WERR are shown in Figure 10d,e, respectively. In general, compared with R_Si/N, H_Total had a greater impact on the WERR of the SiN_x:H film.

3.3. Influence of Deposition Conditions on Yield and Quality

For actual production, in addition to the properties, the yield and quality of SiN_x:H films are crucial, as they directly affect the product’s profitability [46,47]. The SiN_x:H films deposited via PECVD in this research can be used in actual production, necessitating considerations of their yield and quality. Therefore, the effects of different deposition conditions on the deposition rate, non-uniformity, and surface roughness of SiN_x:H films were explored as follows.

3.3.1. Deposition Rate

The DR is closely related to the film yield. The effects of different deposition conditions on the DR of SiN:H film were explored and are shown in Figure 11. With an increase in RF power from 1082 to 1300 W, the DR increased from 105.09 to 108.76 Å/s (Figure 11a). The reason for this is that the increase in RF power can cause more molecules to detach from the plasma and increase the plasma collision, thereby improving the DR [48]. On the contrary, with an increase in electrode plate spacing from 17.1 to 20.9 mm, the DR decreased from 112.67 to 101.54 Å/s. A possible reason for this is that the increase in the concentration of byproducts on the wafer surface led to a decrease in the concentration of reactants, as discussed in Section 3.2.3.

The effects of temperature and chamber pressure on the DR are shown in Figure 11b. As the chamber pressure increased, the concentration of reactants increased but the DR decreased. With an increase in chamber pressure from 3.07 to 3.30 Torr, the DR decreased from 110.74 to 108.42 Å/s, which was related to the reduction in the reactant migration capacity caused by increasing chamber pressure. Similar to the chamber pressure, as the temperature increases, the reactant energy and reaction rate increase but the DR decreases [49]. With an increase in temperature from 360 to 440 °C, the DR decreased from 110.77 to 105.74 Å/s. The probable reason for this is that the reactant migration capacity on the surface of the silicon wafer can also be increased by increasing the temperature, which can increase the film density [21].

An increase in the gas flow ratio can increase the reactant concentration. The more reactants involved in the reaction per unit time, the higher the DR. Therefore, as shown in Figure 11c, the DR increased with an increase in the gas flow ratio. With an increase in the gas flow ratio from 2.98 to 3.51, the deposition rate increased from 103.71 to 114.55 Å/s.

3.3.2. Non-Uniformity

The NU reflects the consistency of the film properties on the entire silicon wafer surface and affects the qualified rate of the product. The effects of different deposition conditions on the NU of SiN_x:H films were explored. As can be seen in Figure 12a, as the RF power increases, the NU shows a downward trend, which is related to the increase in the reactant migration capacity caused by increasing the RF power. With an increase in the electrode plate spacing, the NU shows an upward trend, which can be related to the increase in the concentrations of byproducts on the wafer surface and the edge effect of the electric field. As the temperature increases, the reactant migration capacity on the silicon wafer surface increases; thus, the NU decreases. On the contrary, with an increase in chamber pressure, the reactant migration capacity decreases; thus, the NU increases. Meanwhile, as the gas flow ratio increases, the NU increases. In this research, except under large electrode plate spacing and low temperatures, the NU of the SiN_x:H films deposited via PECVD was less than 1.5%, which indicates that these films can be used in actual production.

3.3.3. Surface Roughness

The surface roughness reflects a film’s surface situation, and it is also closely related to the quality of the film and product. The lower the roughness, the smoother the surface of the film. The effects of different deposition conditions on the surface roughness and AFM results are shown in Figure 13 and Figure 14, respectively. Among them, the surface roughness of the film deposited under the 20.9 mm electrode plate spacing is the largest, which can also be related to the increase in the concentration of byproducts on the wafer surface. Under the different conditions, most of the deposited films had surface roughness values of less than 2 nm, which indicates that films deposited via PECVD have excellent flatness and can be used in actual production.

4. Conclusions

In this study, SiN_x:H films were deposited using mature machines under different deposition conditions. The H-related chemical bonds, RI, E_B, stress, WERR, DR, NU, and surface roughness of the films were tested. Based on the test results, we investigated the effects of deposition conditions such as the RF-power, electrode plate spacing, temperature, chamber pressure, and SiH₄:NH₃ gas flow ratio on the H-related chemical bonds, properties, yield, and quality of SiN_x:H films, and the relationships between the H-related chemical bonds and film properties were studied. The conclusions are summarized as follows:

• Under the different deposition conditions, the differences in the RI and E_B can be mainly attributed to the change in R_Si/N. Meanwhile, H_Total had a greater impact on the WERR of the SiNx:H films, and the stress was not entirely dependent on H_Total.
• As R_Si/N increased, the RI showed a linear growth trend. The reason for this is that the relationship between R_Si/N and the Si:N atomic ratio is positive and linear, and the RI increases with an increase in the Si:N atomic ratio. E_B shows an exponential downward trend with an increase in R_Si/N, which is related to the decrease in the bandgap width.
• Excessive electrode plate spacing should be avoided in the development of the deposition process, as it can lead to significant increases in the NU and surface roughness, reducing the product quality.

In conclusion, the study highlights the critical film properties of SiN_x:H for semiconductor applications. The reflective index reflects the differences in the composition and microstructure and directly affects the lithography process. The breakdown field strength is electrically related to IC reliability, while film stress impacts the binding force between film layers and radio frequency electric field distribution. Additionally, the yield and quality of SiN_x:H films are paramount for production efficiency and economic viability. These findings offer valuable insights for the optimization of semiconductor manufacturing processes.