HAL Id: jpa-00223510
https://hal.archives-ouvertes.fr/jpa-00223510
Submitted on 1 Jan 1983
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
INTERACTIONS OF Si (III) SURFACE WITH H2, NH3, SiH4 MULTIPOLAR PLASMAS STUDIED BY
IN SITU ELLIPSOMETRY
Y. Demay, P. Maurel, S. Gourrier
To cite this version:
Y. Demay, P. Maurel, S. Gourrier. INTERACTIONS OF Si (III) SURFACE WITH H2, NH3, SiH4
MULTIPOLAR PLASMAS STUDIED BY IN SITU ELLIPSOMETRY. Journal de Physique Collo-
ques, 1983, 44 (C10), pp.C10-253-C10-256. �10.1051/jphyscol:19831053�. �jpa-00223510�
JOURNAL DE PHYSIQUE
Colloque CIO, supplément au n°12, Tome HH, décembre 1983 page C10-253
INTERACTIONS OF Si (ill) SURFACE WITH H2, N H3, S i H4 M U L T I P O L A R PLASMAS S T U D I E D B Y IN SITU E L L I P S O M E T R Y
Y. Demay, P. Maurel and S. Gourrier
Laboratoires d'Electronique et de Physique Appliquée, 2, Avenue Descartes, 94450 Limeil Brévannes, France
Résumé - Un système ultra vide formé d'une chambre plasma couplée à une chambre d'analyse permet l'étude de l'interaction de différents plasmas multipolaires (NH,-H?-SiH. ) avec la surface (1,1,1) de silicium. Les cinétiques d'interaction peuvent être suivies en temps réel par ellipsométrie "in situ" à 310 nm. Le même appareillage permet d'analyser l'échantillon par ellipsométrie spectroscopique. La chambre d'analyse est équipée d'un spectromètre Auger (CMA) et d'un système RHEED.
Abstract - An ultra high vacuum system consisting in a plasma chamber and an analysis chamber is used to study the interactions of various multipolar plasmas (NH,-H SiH. ) with Si surfaces. The kinetics of interaction can be followed in real time by in situ ellipsometry at 310 nm. Using the same set up the sample can be analyzed in situ by spectroscopic ellipsometry. The analysis chamber is equipped with an Auger spectrometer (CMA) and a RHEED set up.
Introduction - The use of plasma techniques is rapidly expanding in semiconductor industry. The plasmas can be used for etching, deposition or surface cleaning.
However there is a lack of understanding of the physical processes involved in plasma surface interaction.
This is mostly due to the difficulty of introducing analysis techniques in a plasma environment. In contrast ellipsometry is very well suited for such a study. In this paper we present some results describing the interaction of silicon surfaces with several multipolar plasmas.
Experimental
The system consists in a plasma reaction chamber and an analysis chamber (Fig. 1 ) . Fig. 1
(1) reaction chamber (2) UHV liquid N2 cold trap (3) oil diffusion pump C*) analysis chamber (5) Auger spectrometer (6) RHEED
(7) ion pump
(8) spectroscopic ellipsometer The sample can be transfered from one part to another. The plasma reaction chamber is equipped with an oil diffusion pump associated with an UHV liquid H? trap allowing a base pressure in the 10 Torr range and a rapid pump down (2 minutes) from the plasma pressure (5 10 ) to the 10 Torr range. The plasma source is a multipole (hot cathode discharge associated with a magnetic confinement )(1). During every plasma
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19831053
C10-254 JOURNAL DE PHYSIQUE
reaction the sample was at the floating potential (about -20V). The analysis chamber (base pressure 2 161° Torr) is equipped with an Auger CMA spectrometer and a RHEED gun.
Ellipsometer
Figure 2 shows the spectroscopic ellipsometer set up. Special care was taken for the choice of the view ports in order to minimize spurious polarization effects.
Fig. 2
-
(1) Xenon lamp ( 8 ) diaphragm(2) shutter (9-10) Analyzer
( 3 ) optical encoder (11) filter (4) Rotating polarizer (12) lens
( 5 ) diaphragm (13) monochromator
( 6 ) view port (14) photomultiplier
( 7 ) sample
Kinetics measurements are done at
A
= 310 nm to optimize the optical constrast between a dielectric and the silicon substrate.Results and discussion
a - Surface cleaning - After introduction the silicon surface is covered with 3-4 nm of Si02 and often with carbon contamination (less than 1 nm). Exposure to a H2 plasma suppresses the carbon contamination, but does not seem to reduce the surface oxide. In fact the possibility of cleaning a silicon surface with a H2 plasma seems to depend strongly on the type of plasma source (2). With our
system the surface cleaning was achieved by flash heating at 1200 O C , either under ultra high vacuum or in a H plasma (see Fig. 3). The cleaning procedure is assessed by ellipsometry. RHEED s6ows that the surface is not amorphized.
Silicon Oxygen b) - Surface nitridation
Surface nitridation of the sample is carried out by NH3 olasma exDosure. The behaviour is strOngiy dependent on the nature of the surface before nitridation. Figure 4 shows the nitridation of a SiOZ
! I
covered surface followed by insitu ellipsometry. The slope of the ('f,A) curve is
_
A _ I Y _. A _A.--
consistent with the formationof a oxinitride probably by Fig. 3 - Auger electron spectroscopy : silicon ,,itridation of the wafer cleaned hy flash heating either under silicon surface sio, layer.
ultra vacuum ( 5 10 -10 ) Torr or irl a H2 plasma.
DELTR Fig. 4 - Oxinitridation of a SiOZ
)
I
covered Si Surface"1
~ T ' * 1 1 4 >,,
:1
It is important to note that the2$*,i variation is essentially linear
120
' k
indicating that no additional
roughness is created at the dielectric
~ x l n t t v * d e -substrate interface. Most probably
110 c o v c r r d 51 the nitridation does not reach the
2 3 2 4 25 26 PSI interface.
Furthermore oxidation of a silicon surface does not induce additional roughness (ldyer by layer growth), indicating an abrupt interface ( 3 ) . In
constrast the first stages of plasma nitridation of a
clean
surface shows clearly (point A to point B in Fig. 5 ) , the formation of a strongly absorbant film, most probably due to additional roughness formation.1 3 0 l
I
25 26 2 7 PSI
Fig. 5
-
Nitridation of a clean Fig. 6 - The theoretical curve has Si surface been calculated assuming 2 0 A ofSijN4 with 20 % voids ( H incorpora- tion) on a 10 A interface a-Si
The breakpoint B corresponds to a roughness about 8-loti. This is followed by dielectric growth (from B to C) without roughness increase. The dielectric substrate interface propagates into thc sample while keeping the samc roughness.
Nitridation saturates after the growth of approximatively 20 A . Spectroscopic ellipsometry confirms that in this case the dielectric substrate interface is relatively broad (see Fig. 6).
c) - Silicon nitride deposition - The nitrided samples are covered with plasma deposited silicon nitride. This is achieved by introducing silane in the plasma chamber. Figure 7 shows typical deposition kinetics.
This curve can be fitted by assuming some hydrogen incorporation (about 20 %) a feature commonly observed in plasma deposited silicon nitride (4). The depo~ition rate varies according to the total pressure between 10 A/min and 100 Alrnin. Auger spectroscopy (Fig. 8 ) also shows some oxygen (about 10 % ) .
JOURNAL DE PHYSIQUE
Fig. 7
-
Nitride deposition followed Fig. 8 - Auger Electron Spectroscopy by kinetic ellipsometry ( + ) (A = 310nm) of a deposited layeris fitted by deposition of a nan absorbant film ( . ) (index 1.89 : 8 0 % Si3N4 + 20 % voids)
References
(1) - GOURRIER S., MIRCEA A. and BACAL M., Thin Solid Films 65 (1980) 315.
(2) - CHANG R.P.H., CHANG C.C. and DARACK S., 3. Vac. Sci. Technol.,
3
(1982) 45.(3) - ASPNES D.E. and THEETEN 3.B.), 3. Electrochem. Soc.
127
(1980) 1359.( 4 )
