CC BY
17УДК: 537.525
О МЕХАНИЗМАХ ВЛИЯНИЯ КИСЛОРОДА НА ПАРАМЕТРЫ ГАЗОВОЙ ФАЗЫ И КИНЕТИКУ РЕАКТИВНО-ИОННОГО ТРАВЛЕНИЯ КРЕМНИЯ В ПЛАЗМЕ HBr + CI2 + O2
А.М. Ефремов, В.В. Рыбкин, В.Б. Бетелин, K.-H. Kwon
Александр Михайлович Ефремов*, Владимир Владимирович Рыбкин
Ивановский государственный химико-технологический университет, Шереметевский просп., 7, Иваново,
Российская Федерация, 153000
E-mail: efremov@isuct.ru*, rybkin@isuct.ru
Владимир Борисович Бетелин
ФГУ ФНЦ НИИСИ РАН, Нахимовский просп., 36, к.1, Москва, Российская Федерация, 117218 Kwang-Ho Kwon
Plasma Application Lab., Dept. of Instrumentation and Control Engineering, Korea University, 208 Seochang-Dong, Chochiwon, Korea, 339-800
Проведено исследование влияния соотношений компонентов HBr/Ü2 и CI2/O2 в плазмообразующей смеси HBr+CÍ2+O2 на параметры плазмы, стационарные концентрации активных частиц и кинетику травления Si в условиях, характерных для процессов реактивно-ионного травления: давление плазмообразующего газа (p = 10 мтор), вкладываемая мощность (W = 500 Вт) и мощность смещения (Wdc = 200 Вт). Данные по внутренним параметрам плазмы и кинетике плазмохимических процессов были получены при совместном использовании диагностики плазмы зондами Лангмюра и 0-мерной (глобальной) модели плазмы. Было найдено, что варьирование соотношения HBr/O2 при постоянном содержании CI2 в газовой смеси характеризуется более сильным влиянием на стационарный состав плазмы через реакции под действием электронного удара и процессы атомно-молекулярного взаимодействия, а также обеспечивает более широкое изменение суммарной концентрации атомов галогенов в газовой фазе. Было показано также, что варьирование соотношений HBr/O2 и CI2/O2 в сторону увеличения концентрации кислорода сопровождается снижением скорости травления кремния, которое не коррелирует с изменениями плотностей потоков атомов галогенов и энергии ионов. Модельный анализ кинетики травления позволил заключить, что эффективная вероятность гетерогенной реакции Si + Cl/Br зависит от плотности потока окислительных частиц - атомов кислорода и радикалов OH. Причинами такой зависимости могут являть : 1) окисление кремния, приводящее к увеличению пороговой энергии взаимодействия; и 2) окисление галогенидов кремния в менее летучие соединения вида SiBrxOy и SiClxOy, которое обуславливает снижение скорости ионно-стимулированной десорбции продуктов и доли свободных центров адсорбции для атомов Cl/Br.
Ключевые слова: ионизация, диссоциация, поток атомов галогенов, поток атомов кислорода, поток энергии ионов, скорость травления, эффективная вероятность взаимодействия
Для цитирования:
Ефремов А.М., Рыбкин В.В., Бетелин В.Б., Kwon K.-H. О механизмах влияния кислорода на параметры газовой фазы и кинетику реактивно-ионного травления кремния в плазме HBr + Cl2 + O2. Изв. вузов. Химия и хим. технология. 2019. Т. 62. Вып. 10. С. 76-83 For citation:
Efremov A.M., Rybkin V.V., Betelin V.B., Kwon K.-H. On mechanisms of oxygen influence on gas-phase parameters and silicon reactive-ion etching kinetics in HBr + Cl2 + O2 plasma. Izv. Vyssh. Uchebn. Zaved. Khim. Khim. Tekhnol. 2019. V. 62. N 10. P. 76-83
ON MECHANISMS OF OXYGEN INFLUENCE ON GAS-PHASE PARAMETERS AND SILICON REACTIVE-ION ETCHING KINETICS IN HBr + Ch + O2 PLASMA
A.M. Efremov, V.V. Rybkin, V.B. Betelin, K.-H. Kwon
Aleksandr M. Efremov*, Vladimir V. Rybkin
Department of Electronic Devices and Materials Technology, Ivanovo State University of Chemistry and Technology, Sheremetevskiy ave., 7, Ivanovo, 153000, Russia E-mail: efremov@isuct.ru*, rybkin@isuct.ru
Vladimir B. Betelin
SRISA/NIISI 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
The effects of both HBr/O2 and CWO2 mixing ratios in HBr+Ch+O2 gas mixture on plasma parameters, steady-state densities of active species and Si etching kinetics were studied under the typical conditions of reactive ion etching process: total gas pressure (p = 10 mTorr), input power (W = 500 W), bias power (Wdc = 200 W). The data on internal plasma parameters and plasma chemistry were obtained using a combination of Langmuir probe diagnostics and 0-dimensional (global) plasma modeling. It was found that the variation in HBr/O2 mixing ratio at constant Cl2 fraction in a feed gas is characterized by the stronger impact on the steady-state plasma composition through both electron-impact and atom-molecular reaction kinetics as well as allows one to obtain the wider change in the total halogen atom density. It was shown that changes in both HBr/O2 and Cl2/O2 mixing ratios toward O2-rich plasmas lowers the Si etching rate that exhibits no evident correlations with total halogen atom flux and ion energy flux. The model-based analysis of Si etching kinetics allowed one to conclude that the effective reaction probability for Si + Cl/Br heterogeneous reaction depends on the flux of oxidative species - oxygen atoms and OH radicals. The reasons may be 1) the oxidation of silicon resulting in higher reaction threshold energy; and 2) the decreasing fraction offree adsorption sites for Cl/Br atoms due to the oxidation of reaction products into the lower volatile SiBrxOy and SiClxOy compounds.
Key words: ionization, dissociation, halogen atom flux, oxygen atom flux, ion energy flux, etching rate, effective reaction probability
be desorbed only by ion bombardment. Another positive feature of HBr-containing plasmas is the obtaining
Chlorine- and bromine-containing gases (and of higher etching selectivity for Si and organic photo-
namely, Ch and HBr) play the remarkable role in mod- resist mask. This effects probably follows from the
ern micro- and nano-electronics technology being used graphitization and the cross-linking of polymer surface
for the dry (plasma-assisted) etching of various inor- due to the ultra-violet irradiation from excited HBr
ganic materials. The main application fields for Ch- molecules [8].
and HBr-based plasmas are 1) the patterning of semi- Until now, there were several studies devoted
conductors, metals and metal oxides which form non- to investigations of Si etching characteristics and
or very low-volatile reaction products being treated in mechanisms in Ch- and HBr-based plasmas. The re-
the fluorine-containing environments [1-3]; and 2) the sults of these works may be summarized as follows:
highly-anisotropic etching of both mono- and poly-Si - Effective probabilities for Si + Cl and Si + Br
[3-5]. The last phenomenon is caused by the much reactions have no principal difference under the typical
lower probability of spontaneous chemical reaction in reactive-ion etching conditions. Such conclusion is
the Si + Cl/Br system compared with that for Si + F [6, 7] supported by several experimental facts, and namely
due to the worse penetration of Cl/Br atoms inside the by that: 1) differences in the halogenation degrees for
lattice. That is why, the etching of Si in Ch- and HBr- Si in Ch and HBr plasmas correlate with differences in
containing plasmas is accompanied by the formation of corresponding atom size [9]; 2) silicon etching yields
non-saturated SiClx and SiBrx compounds which may for Cl+, Br+ and HBr+ ions are quite close [10]; and 3)
the variation of Ch/HBr mixing ratio at constant ion bombardment energy leads to the neatly constant Si etching yield [6].
- Silicon etching rates in HBr plasma are lower than those in Cl2 plasma under one and the same processing conditions [5]. Such difference is in agreement with differences in volume densities and fluxes of halogen atoms in corresponding gas systems [9, 11, 12]. Also, the addition of O2 to HBr or Ch slows Si etching rate, but allows one to obtain the more anisotropic etching [13, 14]. The last effect is normally associated with the passivation of sidewalls by low volatile oxy-chlo-rides and bromides.
In our previous works [15, 16], we have performed the combined (experimental and model-based) study to analyze plasma chemistry and Si etching kinetics in the ternary HBr+Cl2+O2 gas mixture. The most principal results were that 1) the composition of a feed gas influences the steady-state densities of neutral species through both electron-impact kinetics and gas-phase processes involving O, O(:D) and OH species; and 2) under the typical conditions of reactive-ion etching process, the silicon etching kinetics corresponds to the neutral-flux-limited regime as well as is affected by the oxidation of SiClx and SiBrx into lower volatile compounds. These findings clearly indicate that the O2 content in a feed gas represent an additional factor to adjust both gas-phase and heterogeneous chemistries in order to optimize output process characteristics. In this work, we attempted to compare the effects of HBr/O2 and Q2/O2 mixing ratios on plasma parameters, steady-state densities of active specie and silicon etching kinetics. The main idea was to demonstrate the ability of HBr+Ch+O2 gas system to control both gas-phase and heterogeneous chemistries by the gas mixing ratios only.
EXPERIMENTAL AND MODELING DETAILS
Experiments were performed in the planar inductively coupled plasma (ICP) reactor [15, 16]. Plasma was excited using a 13.56 MHz RF power supply connected to a flat copper coil on the top side of the chamber. Another 12.56 MHz RF generator powered the bottom electrode in order to produce the negative dc bias voltage (-Udc) and to control the ion bombardment energy. The experiments were carried out at constant gas pressure (p = 10 mTorr), total gas flow rate (q = 40 sccm), input power (W = 500 W), bias power (Wdc= 200 W) while the variable parameter was the composition of feed gas. In one experimental series, the Cl2 flow rate q(Ch) was fixed at 20 sccm while HBr and O2 flow rates were variably set to a combined total
of q(O2) + q(HBr) = 20 sccm. Therefore the fraction of Cl2 y(Cl2) was always 50%, and an increase in q(O2) from 0-10 sccm corresponded to the transition from 50% Cl2 + 50% HBr to 50% Ch + 25% HBr + 25% O2 gas mixture. Another experimental series assumed the constant q(HBr) = 20 sccm as well as the variable flow rates for Cl2 and O2 within q(O2) + q(Ch) = 20 sccm. In this case, the fraction of HBr in a feed gas was always 50% while an increase in q(O2) from 0-10 sccm corresponded to the transition from 50% Ch + 50% HBr to 50% HBr + 25% Ch + 25% O2 gas mixture. Plasma diagnostic provided the data on electron temperature (Te) and ion current density (J+). The latter were extracted from voltage-current curves using the well-known statements of the double Langmuir probe theory [17, 18].
In etching experiments, samples of Si with dimensions of ~ 2x2 cm were located in the middle part of the bottom electrode. The bottom electrode had a built-in water-flow cooling system that allowed one to maintain its temperatures at -17 °C. Accordingly, we assumed constant sample temperatures (TS) in all experimental series as well as ignored the temperature-related effects when analyzing the etching kinetics. The Si etching rates were calculated as R = Ah/x, where Ah is the etched depth measured by the surface profiler Alpha-Step 500 (Tencor), and x = 2 min is the processing time. In order to supply the measurements of Ah, we developed a partial surface masking by the photoresist AZ1512. The thickness of the photoresist layer was -1.5 |im. Preliminary experiments showed that the presence of etched samples in the reactor chamber does not influence the Langmuir probe diagnostics results within the standard experimental error. This allows one to neglect the influence of reaction products on bulk plasma characteristics and thus, to match bulk and heterogeneous subsystems only through fluxes of active species from gas-phase to sample surface.
In order to analyze the influence of HBr/O2 and Cl2/O2 mixing ratios on kinetics and densities of plasma active species, we developed a simplified 0-di-mensional (global) kinetic model operated with volume-averaged plasma parameters. Similar to our previous works [11, 12], the model was based on the simultaneous solution of steady-state kinetic equations with using the experimental data on Te and J+ as the input parameters. The set of chemical reactions was taken from Refs. [15, 16]. Corresponding kinetic schemes have demonstrated an acceptable agreement between model-predicted and measured plasma parameters for pure HBr, Cl2 and O2 plasmas [19-21].
RESULTS AND DISCUSSION
Fig. 1 illustrates the influence of HBr/O2 and Q2/O2 mixing ratios on silicon etching rate (RsO. It can be seen that in increased in y(O2) in both cases lowers the RSi value with different quantitative effects. Particularly, the substitution of HBr for O2 at y(Cl2) = const suppresses Si etching rate by ~ 14 times, in the range of 64.2-4.5 nm/min for 0-25% O2. The influence of Cl2/O2 mixing ratio appears to be much weaker and slightly exceeds ~ 5 times (64.2-12.2 nm/min for 025% O2). The above data clearly demonstrate that the higher O2 content in a feed gas provides the stronger differences in the silicon etching kinetics within one and the same etching mechanism. Previously, several works have mentioned that the quantitative differences in Si etching rates in non-oxygenated Cl2- and HBr-based plasmas are in general agreement with the fluxes of corresponding halogen atoms. In order to understand how this rule works for HBr+Cl2+O2 plasma as well as to figure out the features of Si etching mechanism in the given gas system, the information on plasma parameters and the steady-state plasma composition is mandatory required.
80 Г 1-4
1.2
с 60
S 1.0
!
- 0-8
S 40
Oû с ^ 0.6
J3
£
<D 0.4
м 20
0.2
0 ■ 0.0
10
15
20
, %
25
Fig. 1. Silicon etching rates (1, 2) and effective reaction probabilities for Si + Cl/Br reaction (3, 4) as functions of O2 fraction in a
feed gas at y(Ck) = const (1, 3) and y(HBr) = const (2, 4) Рис. 1. Скорость травления кремния (1, 2) и эффективная вероятность взаимодействия в реакции Si + Cl/Br (3, 4) как функции доли О2 в плазмообразующей смеси при y(Ck) = const (1, 3) и y(HBr) = const (2, 4)
Fig. 2 represents experimental and model-predicted data plasma parameters and densities of charges species. It was found that electron temperature (mean electron energy) exhibits the similar decreasing tendency toward O2-rich plasmas (Te = 2.9-3.3 eV for 025% O2 at y(Ch) = const and 2.9-3.4 eV for 0-25% O2 at y(HBr) = const) due to similar shifts in electron energy losses. Particularly, the substitution of HBr for O2 at y(Cl2) = const changes dominant gas-phase components from HCl to Ch, Cl and O2 (Fig. 3(a)) and causes
a decrease in electron energy losses for vibrational excitation, electronic excitation and ionization. The reasons are 1) higher threshold energies and cross-sections for vibrational excitation of HCl compared with those for Cl2 and O2 [22, 23]; and 2) increasing fraction of atomic species. The substitution of Cl2 for O2 at y(HBr) = const produces similar changes in the densities of both HCl and O2 (Fig. 3(b)) as well as suppresses the density of Cl2 molecules which are characterized by higher electronic excitation and ionization cross-sections (and thus, by higher electron energy losses for corresponding processes) compared with those for O2 [23]. Therefore, an increase in y(O2) in both cases lowers the overall electron energy loss and thus, enriches EEDF by the high-energy electrons.
10"r
Q 1010
10"г
Q 1010-
10
15
20
25
Fig. 2. Electron temperature (1) and densities of charged species (2 -positive ions; 3 - electrons; 4 - negative ions) as functions of O2 fraction in a feed gas at y(Cl2) = const (a) and y(HBr) = const (b) Рис. 2. Температура электронов (1) и концентрации заряженных частиц (2 - положительные ионы; 3 - электроны; 4 - отрицательные ионы) как функции доли О2 в плазмообразующей смеси при y(Cl2) = const (а) и y(HBr) = const (b)
Another important feature of HBr/O2 and Ch/O2 mixing ratios is the different effect on densities of charged species. Particularly, the substitution of HBr for O2 at y(Ch) = const lowers the density of HCl molecules under the condition of [Cl2] + [Cl] > [HCl] for y(O2) > 8% (Fig. 3(a)). Such situation provides an increase in total ionization frequency due to 1) higher ionization rate coefficients for both Cl2 and Cl compared with HCl [22] ; and 2) an increase in all ionization rate coefficients together with increasing Te. This reasonably causes an increase in n+ and ne. Simultaneously, one can obtain the nearly proportional increase
0
5
O, fraction m HBr+CL+O,
%
0
5
O fraction in HBr+CL+O
10"
in both total attachment rate (because of higher attachment rate coefficient for CI2 compared with HCl [22]) and ion-ion recombination frequency for negative ions (because of increasing n+). This results in the nearly constant n- as well as in decreasing plasma electronegativity (n-/ne = 5.1-3.1 for 0-25% O2). Oppositely, the substitution of Cl2 for O2 at y(HBr) = const provides the continuous domination of HCl over other neutral components (Fig. 3(b)) while a decrease in [HCl] (by ~ 2.5 times for 0-25% O2) appears to be stronger than the growth in corresponding ionization rate coefficient (by ~ 1.6 times for 0-25% O2). As such, an increase in O2 fraction in a feed gas leads to monotonically decreasing formation rates for electrons and positive ions as well as causes similar decreasing tendencies for n+ and ne. In addition, the dominant contribution of HCl to the total attachment rate slows the formation of negative ions due to the simultaneous decrease in [HCl] and ne. This leads to a decrease in n- as well as to the nearly constant n_/ne ~ 5.
From Fig. 3, one can conclude that the non-oxygenated 50% HBr + 50% Ch plasma is characterized by 1) the domination of HCl and BrCl over other gasphase components; and 2) the condition [Br] >> [Cl]. When analyzing the kinetics of neutral species, it can be concluded that both phenomena result from the stepwise dissociation of HBr and Br2 through their interaction with Cl atoms in R1: HBr + Cl ^ HCl + Br (ki -1.9x10"" cm3/s) and R2: Br2 + Cl ^ Br + + BrCl (k1 - 1.5x10-10 cm3/s). The frequencies of these processes (k1[Cl] = 75 s-1 and k2[Cl] = 595 s-1) exceed the corresponding values for R3: HBr + e ^ H + Br + e (k3ne = 10 s-1) and R4: Br2 + e ^ 2Br + e (k^ = 100 s-1). Therefore, the mixing of HBr and Cl2 accelerates the consumption of HBr, Br2 and Cl as well as promotes the formation of Br, HCl and BrCl.
The substitution of HBr for O2 at y(Cl2) = const rapidly suppresses densities of HBr, Br2 and BrCl molecules (Fig. 3(a)). The reasons are the acceleration of R3, R4 and R5: BrCl + e ^ Br + Cl + e (due to an increase in Te and ne, see Fig. 2(a)) as well as the stepwise dissociation of above species in reactions involving O and OH. Among the latter, most effective ones are R6: HBr + OH ^ H2O + Br (k6 - 8.0x10-12 cm3/s), R7: "
Br2 + O ^ BrO+Br (k7 - 1.3x10-11 cm3/s), R8: Br2 + OH ^ HOBr + Br (k8 - 3.1 x 10-11 cm3/s) and R9: BrCl + O ^ BrO + Cl (k9 - 2.0x10-11 cm3/s). The density of HCl also shows a decrease toward O2-rich plasmas, but exhibits the weaker change compared with
c1013
с
<D
Q
10*'
bromine-containing species. The first effect is provided by several factors, such as 1) an increase in HCl dissociation frequency through R10: HCl + e ^ H + Cl + e, R11: HCl + O(1D) ^ OH + Cl (kn - 1.0x10-10 cm3/s) and R12: HCl + OH ^ H2O + Cl (k^ - 1.5x10-12 cm3/s); and 2) a decrease HCl formation rate in R1 and R2. The slower change in [HCl] results from an increase in HCl formation rates in R13: Cl2 + H ^ HCl + Cl (k13 -4.0x10-11 cm3/s) and heterogeneous recombination of H and Cl atoms. The density of Cl atoms increases by about an order of magnitude due to the simultaneous increase in their formation rate and a decrease in decay frequency. The mentioned change in Cl atom formation rate is mainly provided by R14: Ch + e ^ 2Cl + e due to the growth of [Q2]. The latter is connected with an increase in formation rate for Cl2 molecules through R15 : BrCl + Cl ^ Cl2 + Br (k:5 - 1.5 x 10-11 cm3/s) and heterogeneous recombination. A decrease in Cl atom decay frequency is mainly connected with R2 and R15 due to decreasing densities of Br2 and BrCl species. The weakly decreasing density of Br atoms is completely associated with the corresponding change in their total formation rate. The latter is produced by the opposite behaviors of R2 and R5 from one side and R15 from the other side.
10"
a)
■ \ . HCl ___Cl2-
BrCl Cl / X -O2'
"ho\
1 V / ''' HOB2 \
' i Br2 ' BrO \
■ HBr :/ \ cio .Л , OH'
. :âH2 ■ .¡АЛ. i /',-''_..HOCl
Л о
10"
10"
10 15 20 25
10 15 20 25
Fig. 3. Steady-state densities of neutral species in HBr+Ch+O2 plasmas with variable HBr/O2 (a) and CI2/O2 (b) mixing ratios Рис. 3. Стационарные концентрации нейтральных частиц в плазме HBr+Ck+O2 с переменными соотношениями компонентов HBr/O2 (a) и CI2/O2 (b)
5
0
5
O fraction in HBr+Cl2+O
2
The substitution of Ch for O2 at y(HBr) = const provides the nearly constants electron-impact dissociation frequencies for molecular species due to an opposite change in Te and ne (Fig. 2(b)). As such, the weakly decreasing densities of HBr, Br2, BrCl and HCl toward O2-rich plasmas (Fig. 3(b)) result only from the effects of gas-phase atom-molecular processes R6-R9, R11 and R12. In contrast to the previous case, the relative contributions of stepwise dissociation processes to total dissociation rates for above species appear to be much weaker. The reasons are 1) the weaker influence of Cl2/O2 mixing ratio on the electron-impact kinetics of R16: O2 + e ^ 2O + e and R17: O2 + e ^ O + O(1D) + e that limits the formation rates for O atoms and OH radicals; and 2) the lower rate coefficients for Cl2 + O/OH reactions compared with those for HBr + O/OH and Br2 + O/OH [23]. The last feature means also the gas system with y(HBr) = const provides higher conversion rates of O and OH into HOBr and BrO and thus, is characterized by lower densities of ox-idant species compared with the case of y(Cl2) = const. The density of Cl atoms shows the weak decrease toward O2-rich plasmas due to a decrease in their formation rates in R5 and R14. The evident reason for such effect is the decreasing densities of source species. The similar behavior of [Br] is also caused by the change in Br atom formation kinetics. Particularly, one can obtain a decrease in Br atom formation rates through R2, R5 and R15 due to the corresponding changes in [Br2] and [BrCl].
The above data allow one to conclude that that change in HBr/O2 and Cl2/O2 mixing ratios leads to quite similar change in [Br] as well as exhibits the different effect on [Cl]. As a result, the gas system with y(Cl2) = const provides the wider adjustment for the total density of chemically active species. In previous works [6, 9, 10], one can find enough arguments that partial probabilities for heterogeneous reactions Si + Cl and Si + Br are quite close. Therefore, the measured Si etching rate may be represented as RSi = Yr(Tci + TbO, where yr is the effective reaction probability, and Tci + rBr is the total flux of halogen atoms to the etched surface. The substitution of HBr for O2 at y(Ch) = const provides an increase in Tci + TBr (4.0x 1017-8.2x 1017 cm"2s"1, or by ~ 2 times for 0-25% O2) due to the rapid increase in Ta. Accordingly, the combination of decreasing etching rate and increasing flux of chemically active species corresponds to the rapidly decreasing yr toward O2-rich plasmas (1.3x10_2-4.6x10~4, or by ~ 30 times for 0-25% O2, see Fig. 1). The substitution of Ch for O2 at y(HBr) = const lowers the Tci + TBr value (4.0x1017-2.7x1017 cm_2s_1, or by - 1.5 times for 0-25%
O2) due to the corresponding changes in partial densities and fluxes. However, the faster change in RSi compared with Tel + TBr toward O2-rich plasmas also corresponds to decreasing yr (1.3x10_2-3.8x10~3, or by - 4 times for 0-25% O2, see Fig. 1). From Refs. [5, 18], it can be understood that the effective probability of ion-assisted chemical reaction at nearly constant surface temperature normally correlates with the ion bombardment intensity. The latter may be characterized by the parameter (MiSi)1/2T+ [24-26], where Mi is the effective ion mass, Si = -e(Udc + Uf) is the ion bombardment energy, -Uf « 0.5Teln(me/2.3m+) is the floating potential, and T+ is the flux of positive ions. It was found that the substitution of HBr for O2 at y(Ch) = const decreases both - Udc (453-429 V for 0-25% O2, see Fig. 4) and ion bombardment energy because of eUdc >> eUf. However, this change is overcompensated by increasing T+ (4.6x1015-6.1x1015 cm_2s_1 for 0-25% O2), so that the parameter (MiSi)1/2T+ indicates the intensification of ion bombardment toward O2-rich plasmas (Fig. 4). The substitution of Cl2 for O2 at y(HBr) = const causes very similar change in - Udc (453-425 V for 0-25% O2) but results in slightly decreasing ion flux in the range of 4.6x1015-4.1x1015 cm_2s_1. Accordingly, the monotonic decrease in (MiSi)1/2T+ takes place (Fig. 4).
9.5 9.0 'S 8.5
>
V 8.0
7.5 , 7.0 6.5
460
450
440
Г >
■a 5
430
420
6.0 L 410
3
10
15
20
25
Fig. 4. Negative dc bias at Wdc = const (1, 2) and the parameter (MiSi)1/2r+ characterizing ion energy flux (3, 4) in HBr+Cl2+O2 plasmas with variable HBr/O2 (1, 3) and Q2/O2 (2, 4) mixing ratios Рис. 4. Отрицательное смещение при Wdc = const (1, 2) и параметр (MiSi)1/2r+, характеризующий плотность потока энергии ионов (3, 4), в плазме HBr+Ch+O2 с переменными соотношениями компонентов HBr/O2 (1, 3) и CI2/O2 (2, 4)
The absence of qualitative (in the case of y(Cl2) = const) and quantitative (in the case of y(HBr) = const) correlation between the changes of yr and (M1s1)1/2r+ allows one to assume that the effective reaction probability is influenced by an additional factor connected with the chemistry of neutral species. Ac-
0
5
O2 traction in HBr+Cl2+O2,%
cording to previously published works [13-16], this really may be the increasing flux of oxidative species To + Toh (2.5x 1015-6.8x 1016 cm_2s_1 for y(Cl2) = const and 1.3x1015-3.4x1016 cm_2s_1 for y(HBr) = const at 1-25%
02) that influences the heterogeneous stages of the etching process through suppressing yR. The mechanisms for such influence may be connected with 1) the direct oxidation of Si atoms into SiOx through surface defects (broken Si-Si bonds) produced by ion bombardment; and 2) the oxidation of Si + Cl/Br reaction products into lower volatile SiBrxOy and SiClxOy compounds. Obviously, the first pathway lowers yR through increasing reaction threshold for halogen atoms while the second one retards the chemical reaction through decreasing fraction of free adsorption sites for etchant species. For the case of y(Ch) = const, the last suggestion is confirmed by the fact that the growth of To + Toh appears to be faster compared with (MiSi)1/2r+. This corresponds to an increase in surface fraction covered by reaction products.
CONCLUSIONS
In this work, we investigated the influence of O2 on gas-phase parameters and Si etching kinetics in HBr + Cl2 + O2 inductively coupled plasma through the changes in HBr/O2 and O2/O2 mixing ratios. The combination of plasma diagnostics by Langmuir probes and plasma modeling allowed one to conclude that the variation in HBr/O2 mixing ratio 1) is characterized by the stronger impact on both electron-impact and atom-molecular reaction kinetics; 2) provides the wider change in the total halogen atom density and flux; and
3) results in increasing ion bombardment intensity toward O2-rich plasmas. The experiments showed that and increase in O2 fraction in a feed gas in both cases lowers the Si etching rate and effective reaction probability for Si + Cl/Br heterogeneous reaction. The last effect may be related to 1) the oxidation of Si by O and OH species resulting in higher reaction threshold energy; and 2) the oxidation of reaction products into the lower volatile SiBrxOy and SiClxOy compounds resulting in decreasing fraction of free adsorption sites for etchant species.
The publication was carried out as part of the state assignment of the Federal State Institution Scientific Center for Research and Development of the Russian Academy of Sciences (fundamental scientific research) on the topic No. 0065-2019-0006 "Fundamental and applied research in the field of subwave holographic lithography, physicochemical etching processes of 3D nanometer dielectric structures for the development of critical technologies for the production of electronic components ".
Публикация выполнена в рамках государственного задания ФГУ ФНЦ НИИСИ РАН (проведение фундаментальных научных исследований) по теме № 0065-2019-0006 «Фундаментальные и прикладные исследования в области субволновой голографической литографии, физико-химических процессов травления 3D нанометровых диэлектрических структур для развития критических технологий производства ЭКБ».
REFERENCES ЛИТЕРАТУРА
1. Wolf S., Tauber R.N. Silicon Processing for the VLSI Era. V. 1. Process Technology. New York: Lattice Press. 2000. 416 p.
2. Advanced plasma processing technology. New York: John Wiley & Sons Inc. 2008. 479 p.
3. Nojiri K. Dry etching technology for semiconductors. Tokyo: Springer International Publishing. 2015. 116 p.
4. Donnelly V.M., Kornblit A. Plasma etching: Yesterday, today, and tomorrow. J. Vac. Sci. Technol. 2013. V. 31. P. 050825-48. DOI: 10.1116/1.4819316.
5. Jin W., 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. 2002. V. 20. P. 2106-2114. DOI: 10.1116/1.1517993.
6. Vitale S.A., Chae H., Sawin H.H. Silicon etching yields in F2, Cl2, Br2, and HBr high density plasmas. J. Vac. Sci. Tech-nol. 2001. V. 19. P. 2197-2206. DOI: 10.1116/1.1378077.
7. Belen R.J., Gomez S., Kiehlbauch M., Aydil E.S. Feature scale model of Si etching in SF6/O2/HBr plasm and comparison with experiments. J. Vac. Sci. Technol. A. 2006. V. 24. P. 350-361. DOI: 10.1116/1.2173268.
8. Pargon E., Menguelti K., Martin M., Bazin A., Chaix-Plu-chery O., Sourd C., Derrough S., Lill T., Joubert O. Mechanisms involved in HBr and Ar cure plasma treatments applied to 193 nm photoresists. J. Appl. Phys. 2009. V. 105. P. 094902. DOI: 10.1063/1.3116504.
9. Cheng C.C., Guinn K.V., Herman IP., Donnelly V.M Competitive halogenation of silicon surfaces in HBr/Cl2 plasmas studied with xray photoelectron spectroscopy and in situ, realtime, pulsed laserinduced thermal desorption. J. Vac. Sci. Technol. A. 1995. V. 13. P. 1970-1976. DOI: 10.1116/1.579638.
10. Ito T., Karahashi K., Kang S.-Y., Hamaguchi S. Evaluation of Si etching yields by Cl+ Br+ and HBr+ ion irradiation. J. Phys.: Conf. ser. 2010. V. 232. P. 012021. DOI: 10.1088/1742-6596/232/1/012021.
11. 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(2). P. 259-271. DOI: 10.1007/s11090-010-9279-7.
12. 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.
13. 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(2). P. S42-S52. DOI: 10.1088/0963-0252/14/2/S06.
14. 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. Mater. Sci. Semicond. Proc. 2007. V. 10(1). P. 41-48. DOI: 10.1016/j.mssp.2006.08.027.
15. Lee B. J., Efremov A., Kim J., Kim C., Kwon K.-H. Peculiarities of Si and SiO2 etching kinetics in HBr+Cl2+O2 inductively coupled plasma. Plasma Chem. Plasma Proc. 2019. V. 39. P. 339-358. DOI: 10.1007/s11090-018-9943-x.
16. Lee B. J., Efremov A., Kwon K.-H. Plasma parameters, gas-phase chemistry and Si/SiO2 etching mechanisms in HBr+Cl2+O2 gas mixture: Effect of HBr/O2 mixing ratio. Vacuum. 2019. V. 163. P.110-118. DOI: 10.1016/j.vacuum.2019.02.014.
17. Shun'ko E.V. Langmuir probe in theory and practice. Boca Raton: Universal Publishers. 2008. 245 p.
18. Lieberman M.A., Lichtenberg A.J. Principles of plasma discharges and materials processing. New York: John Wiley & Sons Inc. 2005. 730 p.
19. 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(12). P. D777-D782. DOI: 10.1149/1.2993160.
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(5). P. H574-H579. DOI: 10.1149/1.3362943.
21. 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(15). P. 3272-3284. DOI: 10.1088/00223727/39/15/009.
22. Chistophorou L.G., Olthoff J.K Fundamental electron interactions with plasma processing gases. New York: Springer Science+Business Media. 2004. 780 p.
23. NIST Chemical Kinetics Database. https://kinet-ics.nist.gov/kinetics/welcome.jsp (15.04.2019).
24. Lee J., Efremov A., Yeom G.Y., Lim N., Kwon K.-H. Application of Si and SiO2 etching mechanisms in CF4/C4F8/Ar inductively coupled plasmas for nanoscale patterns. J. Na-nosci. Nanotechnol. 2015. V. 10. P. 8340-8347. DOI: 10.1166/jnn.2015.11256.
25. Efremov A., Lee J., Kim J. On the control of plasma parameters and active species kinetics in CF4+O2+Ar gas mixture by CF4/O2 and O2/Ar mixing ratios. Plasma Chem. Plasma Process. 2017. V. 37. P. 1445-1462. DOI: 10.1007/s11090-017-9820-z.
26. 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. P. 1344-1351. DOI: 10.1109/TPS.2004.828413.
Поступила в редакцию 16.03.2019 Принята к опубликованию 25.07.2019
Received 16.03.2019 Accepted 25.07.2019
