Plasma Etching
`An Introduction
`
`Edited by
`
`Dennis M. Manos
`Plasma Physics Laboratory
`Princeton University
`Princeton, New Jersey
`
`Daniel L. Flamm
`AT & T Bell Laboratories
`Murray Hill, New Jersey
`
`@
`
`Academic Press
`San Diego New York Boston
`London Sydney Tokyo Toronto
`
`Page 1 of 11
`
`Samsung Exhibit 1017
`
`

`

`This book is printed on acid-free paper. @
`
`Copyright © 1989 by Academic Press
`All rights reserved.
`No part of this publication may be reproduced or
`transmitted in any form or by any means, electronic
`or mechanical, including photocopy, recording, or
`any information storage and retrieval system, without
`permission in writing from the publisher.
`
`ACADEMIC PRESS
`A Division of Harcourt Brace & Company
`525 B Street, Suite 1900
`San Diego, California 92 101 -4495
`
`United Kingdom Edition published by
`ACADEMIC PRESS INC. (LONDON) LTD.
`24-28 Oval Road, London NWl 7DX
`
`Library of Congress Cataloging-in-Publication Data
`
`Plasma etching.
`(Plasma: materials interactions)
`Bibliography: p.
`Includes index.
`I. Manos, Dennis M.
`1. Plasma etching.
`Daniel L.
`III. Series: Plasma.
`TA2020.P5 1988 621.044 87-37419
`ISBN 0-12-469370-9
`
`Alkaline paper
`
`II. Flamm,
`
`PRINTED IN THE UNITED STATES OF AMERICA
`97 EB 98 7 6
`
`Page 2 of 11
`
`

`

`Contents
`
`Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`1 Plasma Etching Technology-An Overview
`Daniel L. Flamm and G. K. Herb
`
`. . . . . . . . . . . . . . .
`
`I.
`Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`II. What Is a Plasma?
`. . . . . . . . . . . . . . . . . . . . . . . . . . . .
`III. Processes in a Plasma . . . . . . . . . . . . . . . . . . . . . . . . . .
`IV.
`Process Requirements and Examples. . . . . . . . . . . . . . . .
`V.
`Plasma Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`VI. Etching Endpoint Detection . . . . . . . . . . . . . . . . . . . . . .
`Vil. Device Damage from the Plasma . . . . . . . . . . . . . . . . . .
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`ix
`
`xi
`
`1
`
`2
`14
`20
`42
`64
`73
`82
`87
`
`2
`
`Introduction to Plasma Chemistry
`Daniel L. Flamm
`
`. . . . . . . . . . . . . . . . . . . . . .
`
`91
`
`I.
`Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`II. How Plasma Etching Takes Place . . . . . . . . . . . . . . . . . .
`III. Etching Characteristics and Variables . . . . . . . . . . . . . . .
`IV. Etching Silicon in Fluorine Atom Based Plasmas . . . . . . .
`V.
`The Loading Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`VI. The Role of Gas Additives. . . . . . . . . . . . . . . . . . . . . . .
`VII. Chlorine Plasma Etching . . . . . . . . . . . . . . . . . . . . . . . .
`VIII. Etchant-Unsaturate Concepts
`. . . . . . . . . . . . . . . . . . . .
`IX. Etching Silicon Oxide in Unsaturated and
`Fluorine-Rich Plasmas
`. . . . . . . . . . . . . . . . . . . . . . . . .
`Silicon Nitride Etching . . . . . . . . . . . . . . . . . . . . . . . . .
`X.
`XI. Oxygen Plasma Etching of Resists
`. . . . . . . . . . . . . . . . .
`XII. 111-V Etching Chemistries and Mechanisms . . . . . . . . . . .
`Acknowledgement
`. . . . . . . . . . . . . . . . . . . . . . . . . . . .
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`92
`92
`99
`131
`138
`144
`146
`155
`
`159
`165
`167
`170
`178
`178
`
`v
`
`Page 3 of 11
`
`

`

`vi
`
`Contents
`
`3 An Introduction to Plasma Physics for Materials Processing . . . .
`Samuel A. Cohen
`
`Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`I.
`The Plasma State . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`II.
`III. Single-Particle Motion . . . . . . . . . . . . . . . . . . . . . . . . . .
`IV. Plasma Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`V.
`Discharge Initiation . . . . . . . . . . . . . . . . . . . . . . . . . . .
`VI. An Application- The Planar Magnetron. . . . . . . . . . . . .
`Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`4 Diagnostics of Plasmas for Materials Processing . . . . . . . . . . . .
`D. M. Manos and H. F. Dylla
`
`I.
`II.
`III.
`IV.
`v.
`VI.
`VII.
`Vlll.
`
`Introduction ........ . ........ . ............. . .
`Electrostatic (Langmuir) Probes .... ... ........... .
`Microwave Interferometry ...................... .
`Impedance Analysis ............ . . .. ....... . .. .
`Mass Spectrometry of Plasmas . . . . . . . . . . . . . . . . . . ..
`Emission Spectroscopy ... . .. ............... ... .
`Fluorescence . . . . . . . . . . . . . . . . ... . .... .... .. . .
`Summary .................................. .
`References . . . . . . . . . . . . . . . . . . . . . . . . .... .. . .. .
`
`5 Plasma Etch Equipment and Technology . . . . . . . . . . . . . . . . . .
`Alan R. Reinberg
`
`Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`I.
`II. Classification of Etch Equipment . . . . . . . . . . . . . . . . . .
`III. Process Control
`. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`IV. Etching Methods- Films
`. . . . . . . . . . . . . . . . . . . . . . .
`V.
`Etching Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . .
`VI. Other Materials
`. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`VII. Discovering and Characterizing Processes . . . . . . . . . . . .
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`6
`
`Ion Beam Etching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`James M. E. Harper
`
`Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`I.
`Broad-Beam Ion Sources . . . . . . . . . . . . . . . . . . . . . . . .
`II.
`Inert Ion Beam Etching . . . . . . . . . . . . . . . . . . . . . . . . .
`III.
`IV. Reactive Ion Beam Etching . . . . . . . . . . . . . . . . . . . . . .
`
`185
`
`186
`187
`193
`205
`241
`248
`258
`258
`
`259
`
`260
`261
`289
`297
`305
`31 2
`325
`331
`332
`
`339
`
`340
`341
`354
`359
`363
`384
`385
`387
`
`391
`
`391
`392
`404
`411
`
`Page 4 of 11
`
`

`

`Contents
`
`V.
`
`. . . . .
`Ion Etching Combined with Growth or Deposition
`Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`7 Safety, Health, and Engineering Considerations for Plasma
`Processing
`G. K. Herb
`
`Legislated Safety Obligations . . . . . . . . . . . . . . . . . . . . .
`I.
`Response to Safety Legislation . . . . . . . . . . . . . . . . . . . .
`II.
`III. Plasma Process Chemistry . . . . . . . . . . . . . . . . . . . . . . .
`IV. Characteristics of Cylinder Gases . . . . . . . . . . . . . . . . . .
`V. Hazardous Gas Monitoring Instruments . . . . . . . . . . . . .
`VI. Checking the Workplace for Hazardous Chemicals . . . . . .
`VII. A Hazardous Aluminum Plasma Etch Process . . . . . . . . .
`VIII. Plasma Process System Configuration . . . . . . . . . . . . . . .
`IX. Recommendations
`. . . . . . . . . . . . . . . . . . . . . . . . . . . .
`Appendix A: Comparison of Federal and State
`. . . . . . . .
`Occupational Safety and Health Requirements
`Appendix B: Characteristics of Common Plasma
`Etch Feeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
`
`Index
`
`vii
`
`418
`421
`421
`
`425
`
`426
`429
`434
`437
`443
`445
`446
`451
`462
`
`463
`
`465
`468
`
`471
`
`Page 5 of 11
`
`

`

`Introduction to Plasma Chemistry
`
`117
`
`will have a rate law that is second or third order, r = k 2n~F2 (ex P 2) (it will
`be third order at pressures where M appears in the rate expression).
`Similarly, the effective kinetics of sequential reactions leading to growth of
`trimers (C3F6 ) in this example would be still higher order [42]. This is
`symptomatic of a general tendency to favor oligomer and polymer growth
`relative to etching, when the pressure increases in a fixed mixture contain(cid:173)
`ing F atoms and CF2 radicals. Similarly, with decreasing pressure, the
`etching rate in this simplified example will rise relative to the rate of
`oligomer and film formation.
`Increasing pressure and decreasing temperatures increase the surface
`concentration of physisorbed species according
`to
`their adsorption
`isotherms, and chemical etching rates increase in step with the surface
`concentration of etchant. While adsorption effects in plasma etching have
`not been studied, in the closely related low pressure gaseous etching
`(LPGE) [46, 47] they can lead to an apparent "negative activation energy."
`Interesting conditions were found in which decreasing temperature led to
`an increase in the rate of silicon etching by XeF2 , ClF3 and other interhalo(cid:173)
`gen compounds, apparently because the surface concentration of active
`species increased faster than the decline in rate constant with temperature
`[46, 47] (see Section 111.1).
`Finally, we note that at fixed composition and mass flow rate, the ratio of
`convective relative to diffusive mass transport (e.g., the Peclet number), is
`constant and both are independent of pressure. Therefore the ratio of mass
`transport rates relative to first order surface reactions will vary as 1/ P so
`that lower pressure tends to overcome local reactant depletion, or mass
`transport limitations in chemical reaction.
`
`E. TEMPERATURE EFFECTS
`
`Temperature, like pressure, has a profound influence on discharge chem(cid:173)
`istry. To be clear, we really should distinguish between gas and surface
`temperatures. However, the gas temperature is a complex function of local
`power input, heat transfer and transport phenomena. Only the surface
`temperature is really controllable. Moreover, for the pressure and flow
`conditions generally encountered in low pressure plasma etching, the ther(cid:173)
`mal boundary layer (e.g., the distance from the surface over which heat
`transfer maintains the gas close to wall temperature) is much thicker than a
`mean-free path, so impinging gas species are already at the surface tempera(cid:173)
`ture.
`In our discussion of kinetics, we said that the rate constants for chemical
`reactions are a function of temperature. Thus temperature has a dominant
`
`Page 6 of 11
`
`

`

`118
`
`Daniel L. Flamm
`
`effect on selectivity, etch rates and the degradation of resist masks. As we
`shall see, the morphology of etched surfaces is also greatly affected by
`temperature. Finally, physisorption and diffusion are sensitive to tempera(cid:173)
`ture, although these topics are beyond the scope of this chapter.
`
`1. Effect of Temperature on Rate Constants
`
`The rate constants for elementary chemical reactions usually vary with
`temperature according to the Arrhenius expression,
`
`(31)
`
`where A is a "pre-exponential" which is weakly dependent on temperature,
`and EA is the "activation energy." The activation energy is the height of the
`energy barrier that reactants must overcome to approach each other and
`combine, (or the energy barrier for dissociation in the case of a single
`decomposing reactant). The exponential term is known as an "Arrhenius
`factor." As an example of this behavior, the etch rates for fluorine atoms
`etching silicon and Si02 are in excellent agreement with the expression
`
`(32)
`
`where the constants, C and EA for Si and Si02 etching are shown in Table
`4. Unlike the expression for the rate constant, Eqn. 31, the reaction rate
`Eqn. 32 depends on a concentration (nF) as given by Eqn. 27. The weak
`pre-exponential T 112 dependence can be understood as follows: The flux of
`F atoms to a surface is nFvF/4 (see Chapter 1), which is proportional to
`r- 112 since n F is inversely proportional to T according to the perfect gas
`
`Table 4 Preexponential factors and activation energy for F atom etching of Si and Si02 •
`The rate equation is ER (A/min) = AnFT112e- t.:,./RT.
`
`FILM
`
`A
`
`EA (kcaljmole)
`
`RATE A/min
`(298 K, nF = 3 X 1015 cm - 3 )
`
`-
`
`-
`
`-
`
`- -- - -
`
`----·---
`
`- -----
`
`2.86 x 10- 12
`0.614 x 10- 12
`- --
`-------- - ---- - - - - - ------- -
`ER(Si)/ER(Si02 ) = 4.66 e!.27/RT
`= 44 (at 298 K)
`
`2250
`55
`
`2.48
`3.76
`
`Page 7 of 11
`
`

`

`Introduction to Plasma Chemistry
`
`119
`
`T (°K)
`50000~~-4~0_0~3_5~0~~3~0_0~~-2_5~0~~
`
`10000
`-;:: 5000
`.E
`' o<t
`~ 1000
`~ 500
`a::
`I
`u
`~ 100
`50
`
`10 ~--~--~--~--~--~10-6
`2.0
`2.5
`3.0
`3.5
`4.0
`4.5
`
`FIGURE 9. Arrhenius plot of silicon and Si02 etch rates, and Si reaction probability with
`F-atoms. The logarithms of these parameters are a linear function of l/T. The etch rates
`shown are based on a plasma F-atom concentration of 3 X 1015 cm - 3. Note that the reaction
`probability, < is defined here as the probability that an impinging silicon atom leaves the
`surface as a silicon fluoride product (no particular product stoichiometry is assumed). Since
`conflicting definitions of < appear in the literature, published values should be interpreted with
`caution.
`
`law, and Vp is proportional to T 112. Hence, if the probability of an atom
`reacting once it is on the surface is proportional to the Arrhenius factor, the
`reaction rate will have the dependence shown. The logarithms of the silicon
`etch rate and probability of an atom reacting when it reaches a silicon
`surface are plotted against l/T in Fig. 9. A straight line in these semiloga(cid:173)
`rithmic coordinates indicates an Arrhenius dependence, and the slope of the
`line gives EA. The weak T 112 factor has almost no effect on the slope in
`these coordinates.
`Incidently, the final silicon-containing product of both Si and Si02
`etching in F- containing discharges is SiF4 • The reaction rates in Table 4
`belong to initial reactions between F atoms and the substrates which form
`intermediate species, not the final product. But they are also the etch rates,
`because the initial step is the slowest rate determining reaction. That is, the
`rate of sequential reactions is determined by the slowest step.
`Since F atom etching rates of both Si and Si02 conform to exponential
`Arrhenius behavior, the selectivity, which is their ratio, is exponential in
`
`Page 8 of 11
`
`

`

`120
`
`Daniel L. Flamm
`
`150 100
`
`T {°C)
`50
`
`0
`
`-50
`
`80
`
`70
`
`60
`
`0
`j:: 50
`<t a:
`w
`~ 40
`a:
`:::c
`u
`1-w
`
`I
`30
`
`25
`
`20
`
`0
`
`2.5
`
`3.5
`3.0
`1000/T (K)
`
`4.0
`
`4.5
`
`FIGURE 10. Selectivity for etching Si over Si02 as a function of 1/T. The decline of
`selectivity with temperature is an effect of the Arrhenius dependence.
`
`1/T as well. Figure 10 shows the selectivity for etching Si relative to Si02
`as a function of 1/T. At room temperature, selectivity is about 44: 1, but if
`the plasma heats the substrate, the selectivity will fall. Conversely, higher
`selectivity can be achieved by cooling the substrates to below room temper(cid:173)
`ature. Notice also that these exponential factors, exp - EA/RT, always ap(cid:173)
`proach unity when RT is much larger than EA. Selectivities decrease with
`temperature and approach the ratio of the Arrhenius pre-exponential fac(cid:173)
`tors -
`in this case 4.7: 1. Surface temperature is an essential variable.
`As we will see, this example is more than an academic exercise. F atoms,
`made from a variety of feed gases, are widely used to etch Si, Si02 and
`
`Page 9 of 11
`
`

`

`Introduction to Plasma Chemistry
`
`121
`
`Si 3N4 • This is usually done by the purely chemical reactions discussed here,
`but anisotropic etching of Si02 with F atoms is also possible, as described
`in Section IX. To study the ideal etching reactions, substrates were exposed
`to dissociated F2 from a discharge in a temperature-controlled cell. How(cid:173)
`ever, F2 is considered too hazardous for process application so other
`sources of F atoms, including NF3 , mixtures of CF4 , C2F6 etc. with 0 2 , and
`SF6/02 are used industrially. The choice between these alternatives is made
`on the basis of economics and side effects on the etching process. The
`plasma chemistry of halocarbon/oxygen mixtures is important in a rich
`variety of etching processes and will be discussed in Section IV.
`While discussing Si etching by F atoms, we should point out that certain
`gaseous fluorine-bearing compounds will react even without a plasma. XeF2
`is probably the best known of these substances, and investigators have
`repeatedly claimed that its reaction with Si follows the same basic mecha(cid:173)
`nisms as F atoms etching. This claim has, however, been repeatedly
`discredited. Several investigators have shown that the reactivity, the rate,
`the intermediates, and the temperature dependence of XeF2 etching silicon
`and Si02 are quite different from F atoms etching. Apparently, the main
`reason XeF2 has been used in so many studies is that it is commercially
`available (in a bottle) and stable. Measuring its reactions doesn't require the
`techniques that are necessary for F atom studies.
`In fact, XeF2 is but one of a family of "plasmaless" low pressure gaseous
`etchants (LPGE) that can be used to etch silicon selectively, making SiF4 as
`a product. Other gases of this type are listed in Table 5, where atomic
`fluorine is included for comparison. Notice that the room temperature etch
`rates and apparent activation energies vary widely, consistent with the
`
`Table 5 Reaction rates for LPGE etching
`of Si by various gases (at room temperature; • means not measured).
`
`Ea
`Reactant
`Kcal/mole
`- - - - - -- - -- --·
`
`Xefi
`BrF3
`F
`IF5
`BrF5
`CIF3
`ClF
`F2
`
`6.1
`
`2.5
`2.5
`*
`4.1
`*
`9.2
`
`Etch Rate
`A./ min-Torr
`
`230,000
`50,000
`9,200
`2,200
`1,500
`1,200
`< 2
`0.3
`
`
`
`Page 10 of 11
`
`

`

`122
`
`Daniel L. Flamm
`
`580
`500
`620 I 540 I 460 420 380
`
`340
`
`300
`
`260
`
`T.K.
`
`R= (µ.-min-1)
`P =(Torr)
`
`!\I
`.....
`'<"'
`f-
`~ 100
`n::
`
`10'---'--~-'-~---'~--''--~'--~~~-'---'""-~-'-~-"-~--'~--''----'
`1.6
`1.8
`2.0
`2.2 2.4
`2.6 2.8 3.0
`3.2
`3.4 3.6
`3.8 4.0
`
`FIGURE 11. Etch rate of (100) silicon by XeF2 as a function of temperature (upper curve
`and data points). The thick solid lines are asymptotes showing a normal Arrhenius behavior at
`high temperature (negative slope) and anomalous positive slope at low temperature which is
`attributed to adsorption control. For comparison, the etch rate by an equivalent pressure of F
`atoms is shown by the lower line ( -
`• -
`• ).
`
`diverse chemical nature of these etchants. Unlike atomic fluorine, none of
`the plasmaless etchants attack Si02 at all-at least within the several
`hundred degree temperature range in which they have been studied. An(cid:173)
`other interesting aspect of their behavior is that many of them exhibit a
`"counter-Arrhenius" behavior at low temperature. Figure 11 shows this
`peculiarity for XeF2 , which contrasts with the temperature dependence of F
`atom etching. It is believed that the increasing etch rate with decreasing
`temperature means that adsorption of XeF2 on the surface is rate(cid:173)
`that is, an increase in the surface concentration of XeF2 with
`limiting-
`falling temperature more than compensates for a decrease in the reactivity.
`Another way of saying this is that the concentration, n A, in Eqn. 27 should
`really be a surface concentration, and that as temperature drops (below the
`minimum in the curve), adsorption makes the surface concentration rise
`faster than the rate, k(T), falls.
`
`
`
`Page 11 of 11
`
`