CC BY
26DOI: 10.6060/ivkkt.20196207.5947 УДК: 537.525
ПАРАМЕТРЫ ПЛАЗМЫ И КИНЕТИКА АКТИВНЫХ ЧАСТИЦ В СМЕСИ HBr + CI2 + O2 А.М. Ефремов, В.Б. Бетелин, К.Х. Квон, Д.Г. Снегирев
Александр Михайлович Ефремов*
Кафедра технологии приборов и материалов электронной техники, Ивановский государственный химико-технологический университет, Шереметевский пр., 7, Иваново, Российская Федерация, 153000 Кафедра естественнонаучных дисциплин, Ивановская пожарно-спасательная академия ГПС МЧС России, пр. Строителей, 33, Иваново, Российская Федерация, 153040 E-mail: efremov@isuct.ru*
Владимир Борисович Бетелин
Федеральный научный центр Научно-исследовательский институт системных исследований РАН, Нахимовский просп., 36, к.1, Москва, Российская Федерация, 117218
Кванг-Хо Квон
Лаборатория применения плазмы, Департамент разработки средств и методов контроля, Университет Корея, 208 Сеочанг-Донг, Чочивон, Республика Корея, 339-800
Дмитрий Геннадьевич Снегирев
Кафедра естественнонаучных дисциплин, Ивановская пожарно-спасательная академия ГПС МЧС России, пр. Строителей, 33, Иваново, Российская Федерация, 153040
В данной работе проведено комбинированное (экспериментальное и теоретическое) исследование характеристик газовой фазы индукционного 13,56 МГц ВЧ-разряда низкого давления в трехкомпонентной смеси HBr + CI2 + O2. Данные по внутренним параметрам плазмы, кинетике плазмохимических процессов и стационарному составу газовой фазы были получены при совместном использовании диагностики плазмы зондами Лангмюра и 0-мерной (глобальной) модели плазмы. Условия эксперимента и моделирования соответствовали постоянным значениям общего давления плазмообразующего газа (p = 10 мтор), вкладываемой мощности (W = 500 Вт), мощности смещения (Wdc = 200 Вт) и содержания кислорода в смеси (y(O2) = 11 %). В качестве варьируемого параметра выступало соотношение начальных концентраций компонентов в паре HBr + CI2, которое изменялось в диапазоне 0 - 89 % CI2. Было найдено, что замещение HBr на CI2 в исследуемом диапазоне условий: 1) приводит к увеличению средней энергии и концентрации электронов; 2) вызывает немонотонное (с максимумом при ~ 45 % CI2) изменение концентрации атомов Br; и 3) обеспечивает рост концентрации атомов О в условиях y(O2) = const. На основе расчетных данных по плотностям потоков активных частиц (атомов Br, Cl и O, положительных ионов) проведен анализ возможных механизмов влияния соотношения начальных концентраций в паре HBr + CI2 на кинетику травления Si и SiO2.
Ключевые слова: параметры плазмы, скорость реакции, поток атомов галогенов, поток энергии ионов, скорости травления Si и SiO2
PLASMA PARAMETERS AND KINETICS OF ACTIVE SPECIES IN HBr + CI2 + O2 GAS MIXTURE A.M. Efremov, V.B. Betelin, K.-H. Kwon, D.G. Snegirev
Aleksandr M. Efremov*
Department of Electronic Devices and Materials Technology, Ivanovo State University of Chemistry and Technology, Sheremetevskiy av., 7, Ivanovo, 153000, Russia
Department of Natural Science Courses, Ivanovo Firefighting and Rescues Academy, Stroiteley av., 33, Ivanovo,
153040, Russia
E-mail: efremov@isuct.ru*
Vladimir B. Betelin
Scientific Research Institute of System Development RAS, Nakhimovsky ave., 36, bld. 1, Moscow 117218, Russia
Kwang-Ho Kwon
Plasma Application Lab., Dept. of Instrumentation and Control Engineering, Korea University, 208 Seochang-Dong, Chochiwon, Korea, 339-800
Dmitriy G. Snegirev
Department of Natural Science Courses, Ivanovo Firefighting and Rescues Academy, Stroiteley av., 33, Ivanovo, 153040, Russia
In this work, we performed the combined (experimental and model-based) study of gasphase plasma characteristics for HBr + Ch + O2 gas mixture under conditions of low-pressure inductive 13.56 MHz discharge. The data on internal plasma parameters, plasma chemistry as well as the steady-state plasma composition were obtained using a combination of Langmuir probe diagnostics and 0-dimensional (global) plasma modeling. Both experimental and modeling procedures were carried out at constant total gas pressure (p = 10 mTorr), input power (W = 500 W), bias power (Wdc = 200 W) and O2 fraction in a feed gas (y(O2) = 11 %). The variable parameter was the HBr + Cl2 mixing ratio, which was changed in the range of 0 - 89 % Ch. It was found that, under the given set of experimental conditions, the substitution of HBr for Ch: 1) results in increasing both mean electron energy and electron density; 2) causes the mon-monotonic (with a maximum at ~ 45 %o Ch) change in Br atom density; and 3) provides an increase in O atom density at y(O2) = const. The possible impacts of HBr + Cl2 mixing ratio on Si and SiO2 etching kinetics were estimated through the analysis of model-predictedfluxes for plasma active species (Br, Cl and O atoms, positive ions).
Key words: plasma parameters, reaction rate, halogen atom flux, ion energy flux, Si and SiO2 etching rates Для цитирования:
Ефремов А.М., Бетелин В.Б., Квон К.Х., Снегирев Д.Г. Параметры плазмы и кинетика активных частиц в смеси HBr + Cl2 + O2. Изв. вузов. Химия и хим. технология. 2019. Т. 62. Вып. 7. С. 72-79 For citation:
Efremov A.M., Betelin V.B., Kwon K.-H., Snegirev D.G. Plasma parameters and kinetics of active species in HBr + Cl2 + O2 gas mixture. Izv. Vyssh. Uchebn. Zaved. Khim. Khim. Tekhnol. 2019. V. 62. N 7. P. 72-79
INTRODUCTION low etching anisotropy of both mono- and poly-silicon.
The last problem may be solved by using chlorine - and
Silicon and silicon dioxide have found numer- bromine-based gas chemistries in a form of Ch- and
ous applications in modern electronic device technol- HBr-containing gas mixtures [6]. There were several
ogy [1-3]. Since most of the applications assume pre- studies have reported on the plasma-assisted etching
cision patterning of the material surface, the develop- characteristics for Si and SiO2 in HBr- and Ch-based
ment of dry etching processes for these materials is an chemistries, for example Refs. [6-12]. When summa-
important task to be solved for achieving advanced de- rizing published data, one should mention following
vice performance. principal issues:
Until now, there were a large number of exper- 1) The probability of spontaneous chemical reimental and theoretical works dealt with etching char- action Si + Cl/Br at typical process temperatures is acteristics for both Si and SiO2 in fluorine-based gas much lower than that for Si + F. This effect is normally chemistries [3-5]. Results of these works provide de- attributed to bigger sizes of Cl/Br atoms that retard tailed data on relationships between processing condi- their penetration inside the lattice of etched material. tions (pressure, input power, and bias power), internal Also, the spontaneous reaction SiO2 + Cl/Br is thermoplasma parameters and heterogeneous stages of the dynamically forbidden. Accordingly, the steady-state etching process. Particularly, it was found that the prin- etching of Si and SiO2 in Cl- or Br-based chemistries cipal disadvantage of all fluorine-based gas chemis- requires ion bombardment in order to destruct the Si-O tries is the spontaneous Si + F reaction that results in
bonds and/or sputter the low volatile non-saturated Si-Clx and SiBrx compounds.
2) Etching rates of both Si and SiO2 in HBr plasma are lower than those in Cl2 plasma under one and the same processing conditions. This fact correlates with differences in volume densities and fluxes of halogen atoms in these gas systems. At the same time, the HBr plasma provides more anisotropies etching together with the better etching selectivity in respect to photoresist masks.
3) The addition of oxygen to HBr or Cl2 results in more anisotropic etching of Si and SiO2, but lowers the absolute etching rates for both materials. This effect is attributed to the formation of lower volatile Si-OxBry and SiOxCly compounds that passivate the sidewalls.
Unfortunately, the most of mentioned works had the purely experimental nature and thus, did not analyze the relationships between gas-phase and heterogeneous chemistries. In addition, there are no studies which consider plasma chemistry and etching kinetics in the three-component HBr+Ch+O2 gas mixture. At the same time, the HBr+Ch+O2 gas system may exhibit the positive features of both bromine- and chlorine-based gas chemistries as well as provide some additional effects due to specific changes in plasma parameters and densities of plasma active species.
In this work, we combined the experimental and model-based approaches to analyze plasma chemistry in HBr+Cl2+O2 inductively coupled plasma. The main goals were 1) to determine how the HBr/Ch mixing ratio influences gas-phase plasma characteristics (electron temperature, energy of ion bombardment, densities, and fluxes of plasma active species); and 2) to establish the gas-phase-related parameters applicable for the analysis of Si and SiO2 etching mechanisms.
EXPERIMENTAL AND MODELING DETAILS
The experiments were performed in the planar inductively coupled plasma (ICP) reactor [13]. The reactor chamber had a cylindrical shape (r = 16 cm, l = 12.8 cm) and was made from anodized aluminum. Plasma was excited using a 13.56 MHz rf power supply connected to a flat copper coil on the top side of the reactor chamber while another 12.56 MHz RF power supply was used to produce the negative dc bias voltage (-Udc) on the bottom electrode. The experiments were conducted at constant gas pressure (p = 10 mTorr), total gas flow rate (q = 45 sccm), input power (W = 500 W) and bias power (Wdc = 200 W). The initial content of each component in the HBr+Ch+O2 gas mixture was adjusted through the corresponding partial flow rates. Particularly, the O2 flow rate was always 5 sccm that corresponded to y(O2) = q(O2)/q = 11%. Accordingly,
the variation of q(Ch) in the range of 0-40 sccm provided y(Cl2) = 0-89% as well as the full substitution of HBr for Cl.
Plasma parameters were measured by double Langmuir probe tool DLP2000 (Plasmart Inc). The treatment of I-V curves aimed at obtaining electron temperature (Te) and ion saturated current density (j+) was based on well-known statements of the double Langmuir probe theory [14, 15]. In order to analyze the chemistry of plasma active species, we developed a simplified zero-dimensional kinetic model with using the data of Te and n+ as input parameters [13, 16, 17]. The electron density (ne) was extracted from the measured n+ using the solution of the steady-state chemical kinetic equation for negative ions together with the quasi-neutrality equation [16, 17]. Dissociative attachment rate coefficients for dominant electronegative components were either directly taken from Refs. [16-18] or estimated using known process cross-sections [19]. The set of chemical reactions included in the model was taken from a series of published works devoted to the modeling of HBr+Ar [13, 20], Ch+Ar [18, 21], O2+Ar [18, 22], Cl2+O2 [18, 23], and HBr+Ch [16] plasmas. The given kinetic schemes (reaction sets with corresponding rate coefficients) have demonstrated an acceptable agreement between model-predicted and measured plasma parameters for HBr, Ch and O2 [18, 20, 21, 24]. Similarly, to Refs. [13, 16], the model used following assumptions: 1) the electron energy distribution function (EEDF) is close to Maxwellian one; 2) the heterogeneous chemistry of atoms and radicals can be described in terms of the conventional first-order recombination kinetics; and 3) The temperature of the neutral ground-state species (Tgas) is independent on the feed gas composition. Since the experimental data on gas temperature were not available in this study, we took Tgas = 600 K as the typical value for the ICP etching reactors with similar geometry under the close range of experimental conditions [10-12].
For the analysis of heterogeneous chemistry, the fluxes of neutral species with the volume density n were calculates as Г « 0.25nuT, where ut is the thermal velocity corresponding to the given Tgas value. The total flux of positive ions was evaluated as Г+ = j+/e. The ion bombardment energy was found as Si = -e(Udc + Uf), where Udc is the negative dc bias on the bottom electrode provided by Wdc, and Uf « 0.5Teln(me/2.3m+) is the floating potential. The effective ion mass m+ was determined using fractions of neutral components with accounting for the differences in corresponding ioniza-tion rates and ion transport coefficients.
RESULTS AND DISCUSSION
The principal features of plasma parameters and gas-phase kinetics in the two-component HBr+Ch gas mixture have been discussed in our previous work [16]. Taking into account that the current work 1) deals with processing conditions which are very close to those used in Ref. [16]; and 2) studies the HBr+Cl2+O2 system with y(O2) = const, it would be reasonable to focus the main attention on the effects resulting from the presence of oxygen.
Fig. 1 illustrates the influence of feed gas composition on measured and model-predicted plasma parameters.
2,8 2,6
> 0}
2,4 2,2 2,0 1,8
[ "I
Ö 0,6 <
0,4 12
0,2 0,4 0,6 0,8
В
О
Ii
y(CL)
b
Fig. 1. Measured (1, 2) and model-predicted (3-6) plasma parameters as functions of CI2 fraction in a feed gas. In Fig. a): 1-elec-tron temperature; 2-ion current density. In Fig. b): 3-total positive ion density; 4-total negative ion density; 5-electron density; 6-relative density of negative ions (n-/ne). The process condition are: y(Ü2) = 11 %, W = 500 W and Wdc = 200 W Рис. 1. Экспериментальные (1, 2) и расчетные (3-6) параметры плазмы в зависимости от содержания Cl2 в исходной смеси. На рис. а): 1-температура электронов; 2-плотность ионного тока. На рис. б): 3-суммарная концентрация положительных ионов; 4-суммарная концентрация отрицательных ионов; 5-концентрация электронов; 6-относительная концентрация отрицательных ионов (n-/ne). Условия процесса: y(Ü2) = 11 %, W = 500 Вт and Wdc = 200 Вт
An increase in y(Cl2) from 0-89% (in fact, the full substitution of HBr for Ch in the HBr + Ch + 11%
O2 gas mixture weak increase in Te (2.5-2.8 eV, see Fig 1(a)) due to the decreasing electron energy loss in inelastic collisions with gas species. The reason is that the higher dissociation degree for Cl2 molecules) causes a compared with HBr provides a higher fraction of atomic species in the Cl2-rich plasmas. Accordingly, since Cl atoms are characterized by higher threshold energies and lower inelastic cross-sections (mainly for electronic excitation and ionization) compared with those for HBr, HCl and Br, the simultaneous increase in y(Cl2) and decrease in y(HBr) enriches EEDF by the high-energy electrons and shifts Te toward higher values. Similar effects have been reported for many binary gas mixtures, including HBr+Ar, Ch+Ar and HBr+ Ch plasmas [16, 20, 21]. A growth in both n+ = 2.5x1010-3.5 x 1010 cm-3 and ne = 2.1 x 109-7.8 x 109 cm-3 (Fig. 1(b)) is the result of increasing total ionization frequency. This is due to the much higher ionization rate coefficient for Ch molecules (~ 1.3x10-9 cm3/s vs. ~ 6.7x10-10 cm3/s for HBr and - 6.8x10-10 cm3/s for Br at Te = 3 eV) and an increase in all ionization rate coefficients together with increasing Te (due to Siz > 3/2Te, where Siz is the threshold energy for ionization). The behavior of the ion current density (j+ = 0.5-0.8 mA/cm2 for 0-89% Ch, see Fig. 1(a)) directly correlates with n+. Finally, a decrease in the n_/ne ratio (10.7-3.5 for 0-89% Ch, see Fig. 1(b)) is the result of a nearly constant total negative ion density n_ « 2.7x1010 cm-3. The last effect is due to the nearly proportional increase in both total attachment rate and ion-ion recombination frequency.
Fig. 2 shows the effect of y(Ch) on the steady-state densities of neutral species. It was found that, under the given set of process conditions, pure HBr plasma keeps all the features of neutral species kinetics mentioned in previous works [13, 16, 20]. These are as follows: 1) reactions R1: HBr + e ^ H + Br + e (k1 -1.6x10-19 cm3/s at Te = 3 eV) and R2: Br2 + e ^ 2Br + e (k2 - 1.2x 10-8 cm3/s at Te = 3 eV) represent the nearly equal sources of Br atoms; 2) the condition k2 >> k3 (where R3: H2 + e ^ 2H + e) as well as the fast decay of H atoms through R4: HBr + H ^ H2 + Br (k4 -1.2x10-11 cm3/s) and R5: Br2 + H ^ HBr + Br (k5 -1.2x10-10 cm3/s) results in [Br] >> [H]; 3) high formation rates for H2 in gas-phase process R4 and Br2 in heterogeneous process R6: 2Br ^ Br2 (k6 - 125 s-1) lead to [H2] « [Br2] « [Br]; and 4) the effective recovery of the original HBr molecules through R5 provides the domination of this species over other neutral components.
0,7
0,5
0,0
1,0
а
12
10
8
10
10
6
4
6
2
10
0,0
0,2
0,4
0,6
0,8
1,0
1011 -
- ' • _l_^-U__u.
1.0 0.0
y(Cl2)
Fig. 2. Steady-state densities of oxygen-free (a) and oxygen-containing (b) neutral species as functions of CI2 fraction in a feed
gas. The process conditions correspond to Fig. 1 Рис. 2. Стационарные концентрации бескислородных (а) и кислородсодержащих (б) нейтральных частиц в зависимости от содержания CI2 в исходной смеси. Условия процесса соответствуют рис. 1
The transition to the 89% HBr + 11% O2 gas mixture lowers the Br atom formation rate in R1 + R2, but supplies several additional channels for the dissociation of HBr (R7-R9) and Br2 (R10-R12) owing to their interactions with O, O(1D), and OH: HBr + O ^ OH + Br (k7 - 3.6x10"13 cm3/s)
HBr + O(1D) ^ OH + Br
(R7) (R8) (R9) (R10) (R11) (R12)
(ks - 1.5x10 cm3/s) HBr + OH ^ H2O + Br (k9 - 8.0x10-12 cm3/s) Br2 + O ^ BrO + Br (kio - 1.3 x10-11 cm3/s) Br2 + O(1D) ^ BrO + Br (kn - 1.3x10-10 cm3/s) Br2 + OH ^ HOBr + Br (k12 - 3.1 x10-11 cm3/s) At the same time, the conditions k?[O] + k9[OH] << kme, k10[O] + k12[OH] << k2ne and [O(1D)] << [O] imply that the efficiency of these channels is much lower compared with those of R1 and R2. In addition, each reaction inside the R10-R12 group liberates only one Br atom while the other one appears to be located in the BrO or HOBr molecule. All these suppress the total Br atom formation rate and result in a weak decrease in Br atom density compared with pure HBr plasma.
The substitution of HBr for Cl2 in the HBr + Cl2 + 11% O2 gas mixture provides the decreasing rates of R1 and R2, but simultaneously introduces new formation pathways for Br atoms through R13: HBr + Cl ^ HCl + Br (k13 - 2.0X10"11 cm3/s) and R14: Br2 + Cl ^ Br + BrCl (k14 - 1.5x10-10 cm3/s). Since the latter have quite high rate coefficients, the fast decay of HBr, and Br2, as well as the rapidly increasing formation rates and densities for both HCl and BrCl are observed. Accordingly, reactions R15: HCl + e ^ H + Cl + e (k15 - 1.2x10-9 cm3/s at Te = 3 eV), R16: BrCl + e ^ Br + Cl + e (k16 - 1.1 x 10-8 cm3/s at Te = 3 eV) and R17: BrCl + Cl ^ Cl2 + Br (k17 - 1.5x10-11 cm3/s) also begin to be an essential source of Br atoms in the range of 20-81% Cl2. Another important feature is that the rate coefficients for R7-R9 and R10-R12 are noticeably higher than those for R18-R20:
Cl2 + O ^ ClO + Cl (k18 - 3.0x10-13 cm3/s) (R18)
Cl2 + O^D) ^ ClO + Cl (k19 - 3.6x10-11 cm3/s) (R19)
Cl2 + OH ^ HOCl + Cl (k20 - 4.0x10-14 cm3/s) (R20)
That is why an increase in y(Cl2) sufficiently reduces the consumption rates for O, O(1D) and OH species, provides the rapid increase in their densities (by - 2.7 times for O, - 5.2 times for O(1D) and - 3.7 times for OH at 0-21% Cl2, see Fig. 2) and accelerates the dissociation of both HBr and Br2 through R7-R9 and R10-R12. All these effects, together with increasing electron impact dissociation frequencies for both HBr (km = 2.0-9.7 s-1 for 0-89% Ck) and Br2 (k2ne = 19-82 s-1 for 0-89% Cl2), results in non-monotonic behavior for both the effective Br atom formation rate and [Br] value. The density of Cl atoms increase monotonically, but slower than expected from the linear change in y(Ch). This is due to the high Cl atom consumption rate through R13, R14 and R17 in HBr-rich plasmas.
The data on plasma parameters and densities of neutral species discusses above allows one to evaluate how the HBr/Cl2 mixing ratio may influence the Si and SiO2 etching kinetics. According to Refs. [17, 25], the rate of ion-assisted heterogeneous chemical reactions may be expressed as R = yrF, where r is the flux of chemically active neutral species, and yr is the effective reaction probability. From several published works, it can be understood that 1) the differences in the halogenation degrees for Si surfaces exposed to pure Cl2 and HBr plasmas are in agreement with the differences in corresponding atom size [26]; 2) the differences in silicon etching rates in pure Cl2 and HBr plasmas [27, 28] are in agreement with the differences in corresponding halogen atom and ion fluxes [7, 13, 16];
4
10
3
2
3) the Si etching yields pure Ch and HBr plasmas at one and the same ion bombardment energy are very close [7]; and 4) the change in Si etching yield with variations of Ch/HBr mixing ratio at constant ion bombardment energy does not exceed the experimental error [7]. Therefore, one can suggest the effective reaction probabilities for Br and Cl atoms with Si surface being quite close. Probably, the same suggestion may be applied for SiO2. As such, one can simply operate with the total chemical active flux TBr + Tci when analyzing the Si and SiO2 etching kinetics. The parameter Yr for ion-assisted chemical reaction is controlled by several ion-surface interaction pathways [25]. In the case of Si and SiO2, these are: 1) the ion-stimulated desorption of reaction products, as the boiling points for both SiCl4 (- 58 °C) and SiBr4 (- 154 °C) [29] exceed the typical process temperature; and 2) the destruction of Si-O bonds, as the corresponding bond strength (- 800 kJ/mol) is much higher for than for Si-Br (358 kJ/mol) Si-Cl (- 416 kJ/mol) [29]. The partial rate of each ion-driven pathway is YsF+, where Ys is the ion-type-averaged process yield. Since Ys has a nature of sputtering yield, it is determined by the momentum transferred from the incident ion in a single collision [5]. Accordingly, the influence of HBr/Ch mixing ratio on the kinetics of both ion-stimulated desorption of reaction products and Si-O bond breaking may be characterized by the parameter (MiSi)1/2r+, where Mi is the effective ion molar mass.
From Fig. 3, it can be seen that the substitution of HBr for Cl2 in the HBr+Cl2+O2 feed gas results in mon-otonically increasing TBr + Tci (6.4x1016-1.2x1018 cm"2s_1 for 0-89% Ch). Also, thought decreases in both -Udc (404-351 V for 0-89% Ch) and ion bombardment energy (420-368 eV for 0-89% Cl2) take place, this tendency is completely overcompensated by increasing ion flux (r+ = 2.9x1015-4.8x1015 cm'V1 for 0-89% Ch). As such, the parameter (MiSi)1/2r+ also exhibits an increase in the range 4.8x1017-6.7x1017 eV1/2cm"2s_1. Though the last fact evidently shows an increase in the ion bombardment efficiency, it does not mean directly the same effect on yr. One can imagine that the behavior yr in HBr+Cl2+O2 gas mixture is influenced by an additional factor connected with the chemistry of O atoms. The mechanisms responsible for the negative impact of O atoms on yr for both Si and SiO2 may be connected with: 1) the formation of Si-O bonds that increases the reaction threshold for F atoms; and 2) the oxidation of reaction products into lower volatile SiBrxOy and SiClxOy compounds. Obviously, the last mechanism decreases the desorption yield for reaction
products and, thus, lowers the fraction of etched surfaces that are available for the adsorption of F atoms. Assuming that both ion energy flux and the O atoms flux have a nearly proportional, but opposite influences on yr, one can roughly characterize their competitive effect by the To/(MiSi)1/2r+ ratio. Figure 3 shows that a change in y(Ch) results in a much faster increase in to (1.1x1014-1.2x1016 cm"2s_1, or by - 110 times for 089% Ch) compared with (MiSi)1/2T+ and thus, provides a monotonically increasing To/(MiSi)1/2r+ ratio toward Cl2-rich plasmas. Therefore, a decrease in yr looks quite probable.
> 6,o
о
5,5
+
S 5,0
300
200
4,5 L 100
■ 1
0*4 0,
0,2 0,4 0,6 0,8 1,0
10-
[ m18L
1018 г
! 101
10'
10'
0,0
0,2
04 yea/,6
0,8
1,0
b
Fig. 3. Fluxes of active species and flux-to-flux ratios as functions of CI2 fraction in a feed gas. In Fig. a): 1-negative dc bias; 2-ion energy flux (MiSi)1/2r+. In Fig. b): 3-total halogen atom flux; 4-oxy-gen atom flux; 5-ro/(MiSi)1/2r+ ratio. The process conditions correspond to Fig. 1 Рис. 3. Потоки активных частиц и отношения потоков в зависимости от содержания Cl2 в исходной смеси. На рис. а): 1-отри-цательное смещение на подложкодержателе; 2-поток энергии ионов (MiSi)1/2r+. На рис. б): 3-суммарный поток атомов галогенов; 4-поток атомов кислорода; 5-отношение To/(MiSi)1/2r+.
Условия процесса соответствуют рис. 1
Summarizing above data, one can conclude that the change in HBr/Ch mixing ratio (as well as in any other input parameter which exhibits the similar influence on formation/decay kinetics for plasma active species) may result in the competitive effect of increasing halogen atoms flux and decreasing effective reaction probability. Probably, such situation disturbs
0,0
3
a
0
0
0
0
■4
0
well-known correlations between input process parameters and both Si and SiO2 etching rates found for F-bases etching chemistries and thus, complicates the optimization of process regimes.
CONCLUSIONS
This work dealt with the effect of HBr/Ch mixing ratio in HBr + Cl2 + 11% O2 gas mixture on gasphase plasma characteristics under conditions of low-pressure inductive 13.56 MHz discharge. The investigation procedure combined plasma diagnostics by Langmuir probes and 0-dimensional (global) plasma modeling. It was found that the substitution of HBr for Cl2: 1) results in increasing mean electron energy, electron density and total positive ion density; 2) causes the non-monotonic (with a maximum at ~ 45% Ch) change in Br atom density; and 3) provides the sufficient increase in O atom density. The net changes in gas-phase plasma characteristics may cause the competitive effect of increasing halogen atoms flux and decreasing effective reaction probabilities for both Si and SiO2. The most realistic mechanism for the decrease in effective reaction probabilities toward Cl2-rich plasmas is connected with the formation of low-volatile SiBrxOy and SiClxOy product layers, which reduces the permittivity of halogen atoms and ions toward the etched surface.
The publication was carried out within the framework of the state assignment of the Federal State Institution FNC NIISIRAS (fundamental research) on subject No. 0065-2019-0006 "Fundamental and applied research in the field of sub wave holographic lithography, physical and chemical etching processes of 3D nanometer dielectric structures for the development of critical technologies for the production of ECB ".
REFERENCES ЛИТЕРАТУРА
1. Rooth J.R. Industrial Plasma Engineering. Philadelphia: IOP Publishing LTD. 2001. 658 p.
2. Plasma Etching Processes for CMOS Devices Realization. London: ISTE Press-Elsevier. 2017. 125 p.
3. Makabe T., Petrovic Z. Plasma electronics: applications in microelectronic device fabrication. New York: Taylor & Francis. 2006. 330 p.
4. Wolf S., Tauber R.N. Silicon Processing for the VLSI Era. V. 1. Process Technology. New York: Lattice Press. 2000. 416 p.
5. Lieberman M.A., Lichtenberg A.J. Principles of plasma discharges and materials processing. New York: John Wiley & Sons Inc. 2005. 730 p. DOI: 10.1002/0471724254.
6. Nojiri K. Dry Etching Technology for Semiconductors. New York: Springer. 2015. 120 p. DOI: 10.1007/978-3-319-10295-5.
7. Vitale S.A., Chae H., Sawin H.H. Silicon etching yields in F2, Cl2, Br2, and HBr high density plasmas. J. Vac. Sci. Technol. A. 2001. V. 19. N 5. P. 2197-2206. DOI: 10.1116/1.1378077.
8. Cunge G., Kogelschatz M., Joubert O., Sadeghi N. Plasmawall interactions during silicon etching processes in high-density HBr/Cl2/O2 plasmas. Plasma Sources Sci. Technol. 2005. V. 14. N 2. P. S42-S52. DOI: 10.1088/0963-0252/14/2/S06.
9. Kim D.K., Kim Y.K., Lee H. A study of the role of HBr and oxygen on the etch selectivity and the post-etch profile in a polysilicon/oxide etch using HBr/O2 based high density plasma for advanced DRAMs. Mat. Sci. Semicon. Proc. 2007. V. 10. N 1. P. 41-48. DOI: 10.1016/j.mssp.2006.08.027.
10. Belen R.J., Gomez S., Kiehlbauch M., Aydil E.S. Feature scale model of Si etching in SF6/O2/HBr plasma and comparison with experiments. J. Vacuum Sci. Technol. A. 2006. V. 24. N 2. P. 350-361. DOI: 10.1116/1.2173268.
11. Tinck S., Boullart W., Bogaerts A. Modeling Cl2/O2/Ar inductively coupled plasmas used for silicon etching: effects of SiO2 chamber wall coating. Plasma Sources Sci. Technol. 2011. V. 20. P. 045012. DOI: 10.1088/0963-0252/20/4/045012.
12. Yeom G. Y., Ono Y., Yamaguchi T. Polysilicon etchback plasma process using HBr, Cl2, and SF6 gas-mixtures for deep-trench isolation. J. Electrochem. Soc. 1992. V. 139. N 2. P. 575-579. DOI: 10.1149/1.2069260.
13. Efremov A., Lee J., Kwon K.H. A comparative study of CF4, Cl2 and HBr + Ar inductively coupled plasmas for dry etching applications. Thin Solid Films. 2017. V. 629. P. 39-48. DOI: 10.1016/j.tsf.2017.03.035.
14. Shun'ko E.V. Langmuir probe in theory and practice. Boca Raton: Universal Publishers. 2008. 245 p.
15. Caneses J.F., Blackwell B. RF compensation of double Langmuir probes: modelling and experiment. Plasma Sources Sci. Technol. 2015. V. 24. P. 035024. DOI: 10.1088/0963-0252/24/3/035024.
16. Efremov A., Kim Y., Lee H.W., Kwon K.H. A comparative study of HBr-Ar and HBr-Cl2 plasma chemistries for dry etch applications. Plasma Chem. Plasma Proc. 2011. V. 31. N 2. P. 259-271. DOI: 10.1007/s11090-010-9279-7.
17. Efremov A.M., Kim D.P., Kim C.I. Simple model for ionassisted etching using Cl2-Ar inductively coupled plasma: Effect of gas mixing ratio. IEEE Trans. Plasma Sci. 2004. V. 32. N 3. P. 1344-1351. DOI: 10.1109/TPS.2004.828413.
18. Hsu C.C., Nierode M.A., Coburn J.W., Graves D.B. Comparison of model and experiment for Ar, Ar/O2 and Ar/O2/Cl2 inductively coupled plasmas. J. Phys. D Appl. Phys. 2006. V. 39. N 15. P. 3272-3284. DOI: 10.1088/0022-3727/39/15/009.
19. Chistophorou L.G., Olthoff J.K Fundamental electron interactions with plasma processing gases. New York: Springer Science+Business Media. 2004. 780 p. DOI: 10.1007/978-14419-8971-0.
20. Kwon K.H., Efremov A., Kim M., Min N.K, Jeong J., Kim
K. A model-based analysis of plasma parameters and composition in HBr/X (X=Ar, He, N2) inductively coupled plasmas. J. Electrochem Soc. 2010. V. 157. N 5. P. H574-H579. DOI: 10.1149/1.3362943.
21. Efremov A., Min N.K, Choi B.G., Baek K.H., Kwon K.H. Model-based analysis of plasma parameters and active species kinetics in Cl2/X (X = Ar, He, N2) inductively coupled plasmas. J. Electrochem. Soc. 2008. V. 155. N 12. P. D777-D782. DOI: 10.1149/1.2993160.
22. Lee B.J., Lee B.J., Efremov A., Yang J.W., Kwon K.H. Etching characteristics and mechanisms of MoS2 2D Crystals in O2/Ar inductively coupled plasma. J. Nanosci. Nanotechnol. 2016. V. 16. N 11. P. 11201-11209. DOI: 10.1166/jnn.2016.13478.
23. Kwon K.H., Efremov A., Yun S.J., Chun L, Kim K Dry etching characteristics of Mo and AkO3 films in O2/Ck/Ar inductively coupled plasmas. Thin Solid Films. 2014. V. 552. P. 105-110. DOI: 10.1016/j.tsf.2013.12.013.
24. Efremov A.M., Kim G.H., Kim J. G., Bogomolov A.V., Kim C.I.
On the applicability of self-consistent global model for the characterization of Ck/Ar inductively coupled plasma. Microelectron. Eng. 2007. V. 84. P.136-143. DOI: 10.1016/j.mee.2006.09.020.
25. Gray D.C., Tepermeister L, Sawin H.H. Phenomenological modeling of ion-enhanced surface kinetics in fluorine-based plasma-etching. J. Vac. Sci. Technol. B. 1993. V. 11. N 4. P. 1243-1257. DOI: 10.1116/1.586925.
26. Cheng C.C., Guinn K.V., Herman LP., Donnelly V.M. Competitive halogenation of silicon surfaces in HBr/Cl2 plasmas studied ray photoelectron-spectroscopy and in-situ, real-time, pulsed laser-induced thermal-desorption. J. Vac. Sci. Technol. A. 1995. V. 13. N 4. P. 1970-1976. DOI: 10.1116/1.579638.
27. Bestwick T.D., Oehrlein G.S. Reactive ion etching of silicon using bromine containing plasmas. J. Vac. Sci. Technol. A. 1990. V. 8. N 3. P. 1696-1701. DOI: 10.1116/1.576832.
28. Jin W.D., Vitale S.A., Sawin H.H. Plasma-surface kinetics and simulation of feature profile evolution in Cl2+HBr etching of polysilicon. J. Vac. Sci. Technol. A. 2002. V. 20. N 6. P. 2106-2114. DOI: 10.1116/1.1517993.
29. Lide D.R. Handbook of chemistry and physics. New York: CRC Press. 2018. 1532 p.
30. Guettari M., Gharbi A. A correspondence between Grun-berg-Nissan constant d' and complex varieties in water/methanol mixture. Phys. Chem. Liq. 2011. V. 49. P. 459-469. DOI: 10.1080/00319101003646546.
31. Herraez J.V., Belda R., Diez O., Herraez V. An equation for the correlation of viscosities of binary mixtures. J. Solution Chem. 2008. V. 37. P. 233. DOI: 10.1007/s10953-007-9226-2.
32. Marczak W., Adamczyk N., L^zniak M. Viscosity of associated mixtures approximated by the Grunberg-Nissan model. Int. J. Thermophys. 2012. V. 33. P. 680-691. DOI: 10.1007/s10765-011-1100-1.
33. Messaadi A., Ouerfelli N., Das D., Hamda H., Hamzaoui A.H. Correspondence between Grunberg-Nissan, Arrhenius and Jouyban-acree parameters for viscosity of isobutyric acid+water binary mixtures from 302.15 to 313.15 K. J. Solution Chem. 2012. V. 41. P. 2186-2208. DOI: 10.1007/s10953-012-9931-3.
Поступила в редакцию (Received) 21.12.2018 Принята к опубликованию (Accepted) 25.03.2019
