The effect of nitrogen doping and heat treatment on electrical resistivity of CVD SiC bulks

With the expansion of chip size, the challenge of achieving uniform etching becomes progressively more formidable. Implementing CVD SiC etching rings enhances etching uniformity effectively and offers notable attributes of high purity and prolonged operational lifespan. Controlling the resistivity of CVD SiC etching rings is essential to cater to diverse processes and equipment requirements. This investigation delves into the impact of nitrogen doping and heat treatment on the resistivity of CVD SiC bulks. Elevated nitrogen doping results in a heightened carrier concentration within CVD SiC. In modest doping cases, the grain boundary barrier height escalates with the doping concentration. However, in instances of higher doping concentrations, the grain boundary barrier diminishes with increasing doping concentration. Following heat treatment, there is a rise in the carrier concentration of the sample. Nonetheless, the surge in sample porosity precipitates a mobility reduction, yielding minimal variance in resistivity before and after heat treatment.


Introduction
In the realm of semiconductor manufacturing, etching assumes a pivotal role.Presently, the predominant technique in use is plasma etching [1,2].This procedure mandates stringent uniformity in the reactant flux across the wafer [3].The imperative to fabricate electronic components across the entire expanse of the wafer underscores the need for achieving near-edge uniform material treatment.As wafer dimensions escalate, ensuring a seamless transition between the chemical and electrical attributes of the wafer periphery and the supporting materials becomes indispensable [4,5].The SiC etching ring is a prevalent support material [6], frequently synthesized through chemical vapor deposition to guarantee a pure etching milieu.
In catering to the requirements of distinct etching processes and equipment, the adjustability of SiC etching ring resistivity is paramount.Furthermore, resistivity significantly influences the lifespan of the etching ring.Presently, investigations into the electrical properties of polycrystalline SiC are predominantly centered on sintered SiC ceramics.Many studies have probed the effects of doping, grain boundary configuration, soluble atoms, SiC crystal structure, porosity, and grain size on the resistivity of sintered SiC [7][8][9][10][11][12][13][14][15].
Among these, nitrogen doping is a prevalent method for controlling resistivity, amassing substantial research attention [16][17][18].Throughout the nitrogen doping process in SiC, N atoms favorably occupy C atom sites.A positive charge is established alongside an additional valence electron upon substituting a C atom with an N atom.Consequently, the charge carriers generated take the form of electrons.Conventionally, donor impurities capable of liberating electrons to generate conductive electrons are categorized as n-type or donor impurities.The process of electron release is termed donor ionization.This process entails the donor impurity undergoing ionization, resulting in the formation of positively charged ions.Simultaneously, the energy emanating from ionization facilitates the transition of surplus valence electrons from a bound state to the conduction band, effectively engendering conductive electrons.Hence, nitrogen doping induces ionization in donor impurities, propelling the donor energy level of SiC closer to the conduction band's edge, thereby rendering SiC conductive [19,20].
However, literature concerning resistivity regulation in CVD polycrystalline SiC remains sparse.This inquiry employs nitrogen as the doping gas to fabricate CVD SiC bulks.It meticulously investigates the influence of nitrogen doping and heat treatment on SiC's microstructure and resistivity, delving into the underlying mechanisms.The outcomes of this research hold promise not only for tailoring the electrical attributes of etching rings but also for broadening the application scope of high-purity CVD SiC.

Experimental procedures 2.1. Material preparation
The preparation of nitrogen-doped SiC bulks was accomplished by utilizing hot wall CVD. Figure 1 presents the schematic diagram of the employed CVD apparatus.The CVD reaction chamber possessed dimensions with a diameter of 1 meter and a height of 1.5 meters.The graphite matrix was suspended on the initial layer of the tooling.Methyltrichlorosilane (MTS) was the precursor, and N2 was the doping gas.The liquid MTS underwent vaporization within the evaporator, facilitated at a temperature of 353 K, and subsequently introduced into the CVD reaction chamber.The ratio of hydrogen (H 2 ) to MTS was consistently maintained at 10:1, while the flow rate of N 2 was meticulously regulated within the range of 0.2 to 20 l min −1 .Following traversal through a cold trap and an active nitrogen filter, the resultant tail gas traversed a sodium hydroxide spray system to eliminate deleterious acidic by-products effectively.After eliminating graphite content from the deposited sample, a heat treatment (HT) was executed within an argon atmosphere at an elevated temperature of 1600 °C for 10 h.The principal deposition parameters employed for generating N-doped SiC bulks are succinctly summarized in table 1.

Coating characterization
The surface morphology and microstructure of the coatings were analyzed using a scanning electron microscopy.The phase composition and orientation were studied by a Smartlab 9 kw x-ray diffractometer (C u Ka radiation, wavelength 0.15406 nm, voltage 40 KV, current 40 mA).The resistivity, mobility, and carrier concentration of the 3C-SiC bulks were measured by an Ecopia Hall Effect Tester (HMS-7000).The porosity is measured using a high-performance, fully automatic mercury intrusion meter (Micromeritics Autopore V 9605).

Results and discussions
3.1.Microstructure characterization Figure 2 illustrates the surface morphology of both untreated and treated samples crafted at distinct nitrogen flow rates.For samples produced under varying nitrogen flow rates, SiC micro morphologies are displayed.The specimens with N 2 flow rate of 0.2 and 20 l min −1 exhibit prototypical fine crystal morphologies, as evidenced in figure 2(a), while figure 2(b) portrays a pyramid-like structure.Interestingly, compared to the previous study [21], doping with increased nitrogen concentration usually induces the development of pyramid morphology, but the observation results in this article are different.The experimental CVD reactor employed in this investigation features an expansive cavity and suboptimal flow field stability.Throughout the fabrication process of samples with N 2 flow rate of 2 l min −1 , the reactant concentration proximate to the sample remains subdued, resulting in a diminished surface nucleation rate.In such instances, the mass transfer process regulates the reaction, thereby giving rise to the pyramid structure observed.After the heat treatment, the sample's grain size Figure 3 shows the XRD spectra of different samples.All of them are typical 3 C Sic spectra without a second phase.Surface defects may cause a small peak at 33.8 °.When the nitrogen flow rate is low (F N2 = 0.2 and 2 l min −1 ), SiC bulks exhibit a strong 〈111〉 preferred orientation.The preferred orientation changes to 〈220〉 when the nitrogen flow rate is high.β-SiC with 〈111〉 orientation has a polar crystal structure containing Si {111} and C {-1-1-1} planes.The N atoms mainly occupy the C position in the silicon carbide lattice.However, during the growth process of β-SiC grains with 〈111〉 orientation, the adsorption of Si atoms by C or N along the silicon end surface is relatively weak.With the increase of N 2 value, the position competition between C and N atoms becomes more intense, which may reduce the growth rate of 〈111〉 orientation, leading to an increased growth rate of 〈220〉 orientation.After heat treatment, the main change is that the preferred orientation of the sample with N 2 flow rate of 20 l min −1 changes to 〈111〉, which is consistent with the characteristics of platelet morphology.

Electrical resistivity 3.2.1. Effect of N 2 flow rate
In the context of polycrystalline materials, the presence of grain boundaries plays a pivotal role in influencing mobility.As outlined in Seto's single-level grain boundary trap model, an abundance of crystal defects and dangling bonds situated within the grain boundary introduce deep-level trap states into the bandgap.Once these trap states capture charge carriers, they become charged themselves, erecting a grain boundary barrier that impedes the transit of charge carriers between adjacent grains.Consequently, the mobility of charge carriers could be improved.This grain boundary barrier is subject to modulation by factors such as doping concentration and grain size.When the doping concentration is low, and only some traps are filled, the grain boundary barrier [22] where q is the electronic charge, L is the grain size, N represents the doping concentration, and ε is the dielectric constant.When the doping concentration is high, and the traps are fully filled, the grain boundary barrier [22] is where Q t is trap density.According to equations (1) and (2), the relationship between the grain boundary barrier and doping concentration is obtained, as shown in figure 4.
The above theory can help us better analyze the results of this study.Figure 5 offers insights into the resistivity, carrier concentration, and mobility of nitrogen-doped SiC blocks.Prior observations have indicated that heightened nitrogen flow rates correspond to increased nitrogen doping levels and concomitant augmented carrier concentrations.Nonetheless, in this study, the carrier concentration was at its nadir when the N 2 flow rate was 2 l min −1 .This finding aligns with the previous morphological analysis, wherein regions proximate to the sample registered lower gas concentrations, consequently reflecting lower nitrogen doping concentrations.Initially, mobility remained nearly invariant before witnessing a notable upswing.
Within the purview of this study, the alteration in mobility is molded by grain size and the height of the grain boundary barrier.A comparative analysis between the sample with N 2 flow rate of 0.2 l min −1 and 2 l/min underscores that the latter exhibits substantially larger grain dimensions and a reduced grain boundary volume fraction, a configuration conducive to improved mobility.However, due to both samples existing in lowconcentration doping, the sample with N2 flow rate of 2 l min −1 displays a lower effective doping concentration alongside larger grain dimensions, consequently manifesting a heightened grain boundary potential barrier that suppresses mobility.As a result, the mobility of samples with N 2 flow rates of 0.2 l min −1 and 2 l min −1 is akin.Conversely, for the sample with N 2 flow rate of 20 l min −1 , the substantial doping concentration markedly reduces the grain boundary barrier, thus engendering heightened mobility.

Effect of heat treatment
Illustrated in figure 5, the application of heat treatment induces a remarkable elevation in the carrier concentration of the sample, registering an increase spanning 1 to 2 orders of magnitude.The pre-existing trend of carrier concentration variation with nitrogen flow rate endures post-heat treatment.Conversely, a substantial decline in mobility is observed.Although the post-heat treatment mobility pattern in response to nitrogen flow rate aligns with the pre-treatment configuration, the disparities in numerical values are marginal.
Throughout the deposition process, nitrogen atoms within the SiC blocks do not achieve complete activation to displace C atoms; some nitrogen atoms persist within atomic interstices.After heat treatment and under specific thermodynamic conditions, atoms can amass adequate energy for rearrangement.Nitrogen atoms that initially remained dormant within atomic interspace become mobile, engendering covalent bonds with silicon atoms and giving rise to additional free electrons, thereby augmenting carrier concentration.
After heat treatment, on the one hand, some atoms in the block can move to low energy positions, forming new grains, reducing the number of structural defects such as dangling bonds and vacancies within the grains, thereby reducing the scattering centers of charge carriers and improving their mobility.On the other hand, after heat treatment, the grain size increases, the number of grain boundaries decreases, and the mobility also increases.However, in this study, the mobility significantly decreased after heat treatment.The definition of mobility [23] is where τ is the average free time of the electron, m is the effective mass of the electron.among them, the value of τ [23] is where P is scattering probability.
The decline in mobility points to an elevation in scattering probability.Surface microstructural examination corroborates this observation, revealing a decrease in surface density following heat treatment.Figure 6 underscores the porosity alteration in samples before and after heat treatment, with the latter exhibiting a substantial surge.This surge in porosity translates to heightened scattering centers and, consequently, a decrease in mobility.

Conclusions
This paper delved into the influence and underlying mechanisms of nitrogen doping and heat treatment on the resistivity of CVD SiC bulks.The primary conclusions of this study are as follows: (1) The resistivity emerges from carrier concentration and mobility.Increased nitrogen doping yields a heightened carrier concentration within CVD SiC.Meanwhile, mobility is chiefly governed by grain size and the potential barriers within grain boundaries.In cases of modest doping, a proportional escalation in doping concentration corresponds to an augmented height of the grain boundary barrier.However, in scenarios characterized by higher doping concentrations, an inverse relationship prevails, causing a reduction in the grain boundary barrier.
(2) After heat treatment, two key developments: carrier concentration increase and grain size expansion.Nevertheless, this augmentation in sample porosity engenders diminished mobility.Consequently, the alteration in resistivity before and after heat treatment remains minimal.However, further research is needed on the impact of heat treatment.

Figure 2 .
Figure 2. SEM images of samples with different nitrogen flow rates before (a), (b), and (c) and after heat treatment (d), (e) and (f).

Figure 3 .
Figure 3. XRD spectra of samples with different nitrogen flow rates before and after heat treatment.

Figure 4 .
Figure 4. Functional dependence of the potential barrier height on doping concentration.

Figure 5 .
Figure 5. Resistivity, carrier concentration, and mobility of samples with different nitrogen flow rates before and after heat treatment.

Figure 6 .
Figure 6.Porosity of samples with different nitrogen flow rates before and after heat treatment.