planarization of microelectronic structures...
TRANSCRIPT
-
Journal of ELECTRICAL ENGINEERING, VOL. 53, NO. 3-4, 2002, 86–90
PLANARIZATION OF MICROELECTRONICSTRUCTURES BY USING POLYIMIDES
Ladislav Matay∗— Róbert Andok
∗∗
In the recent years there has been an increasing demand for new types of polyimides which could planarize topographiesof microelectronic structures by conserving all of their electro-physical properties for a given application. In this work,planarization properties have been examined for the PPID polyimide material. Profilometer technique has been used tomeasure step height changes of isolated lines of various dimensions. A model of local and global planarization has beencompleted. Thermogravimetric analysis (TGA) has been used to evaluate thermal processes in the PPID polyimide.
K e y w o r d s: polyimides, planarization, degree of planarization, RIE, application
1 INTRODUCTION
In the recent years there has been an intensive andrapidly growing interest in spun-on polymer materials ap-plicable as thin isolation or passivation films on semicon-ductor devices, replacing silicon nitride and silicon diox-ide and other commonly used materials. Especially poly-imides have gained their unique popularity in microelec-tronics and, comparing to inorganic dielectrics, one of theadvantages of using polyimides as dielectrics is the abil-ity of planarization of polyimide films over topographiesby solution coating of polyimide [1]. Besides their low di-electric constant (ε < 3.5), so far best achieved degreeof planarization (DOP ) among all presently used mate-
rials, and improved adhesive power, still other superiorproperties of polyimides that have been reported in liter-ature [1, 2] should be noted: insolubility, fine patterning,photo-sensitivity, high short-term resistance, good spin-coating, low water uptake, reduced stress and improvedthermomechanical properties, and relatively low cost.
In general, there are two types of planarization thatplay an important role for lithography — local planariza-tion, which is the ability to planarize closely spaced fea-tures; and global planarization, which is the ability toplanarize topographies across a larger area or the wholewafer [2]. Ideal planarization that would yield a perfectlyflat surface, regardless of the complexity and step varia-tion of the underlying features, can never be achieved inpractice.
Various authors [3–8] have dealt with the problem ofplanarization but quite often only local planarization orglobal planarization has been examined separately. Theaim of this article is to give a summary of achievedgoals for the local and global planarization simultane-ously, while emphasis will be paid to the planarization
properties of the used polyimide material (see more de-tails in the text).
The degree of local planarization, as reported in [3],is determined by the ratio of the step height with thepolyimide (PI) film to the initial height of the metalpattern; or by the slope of the resulting step.
Four different thicknesses are introduced (see Fig. 1):hm is the original step height and hs is the final poly-imide step height, Tbot is the polyimide thickness at thebottom of a step and Ttop is the maximum polyimide filmthickness on top of the step. It is generally less than Tbotprovided that the line feature is small. The difference be-tween the measured step height and the final step height(hm and hs , respectively) gives the change in the poly-mer film thickness (Tbot−Ttop ). The ratio hs/hm is oftenreferred to as the Planarization Constant .
The degree of planarization (DOP ), as defined in [4],is given by the following equation:
DOP = 1 −hshm
, (1)
whilehm − hs = Tbot − Ttop (2)
where, as can be seen in Fig. 1, the differences betweenhm and hs are equal to those between Tbot and Ttop ,respectively. Another factor that is often used in literaturefor defining the DOP is the angle θ of the resulting step,which can be calculated as:
tg θ =2h
w, (3)
where w is the width of a step.
Global planarization, on the other hand, can be definedas the long range effects over a wide patterned area [3]. As
∗Institute of Informatics, Slovak Academy of Sciences, SK-842 37 Bratislava, Slovakia, [email protected]
∗∗Department of Microelectronics, FEI STU, SK-812 19 Bratislava, Slovakia.
ISSN 1335-3632 c© 2002 FEI STU
-
Journal of ELECTRICAL ENGINEERING VOL. 53, NO. 3-4, 2002 87
hs
Tmin
Tmaxq
hm
Substrate
bot
top
T
MESA structure
T
Fig. 1. Degree of planarization for polyimide over a metal step.
Substrate
d(T -h +h )bot m s
dhs
dTbotPolyimide
Substratew x
Tbot
Tbot- h
h
m
m
Polyimide
(a)
(b)
Fig. 2. Wet a) and dry b) polyimide film over positive topographyfeatures. Tbot is the thickness of the coating solution above thesurface of the wafer, hm and w are the height and width of thepositive topography, respectively, hs is the final PI step height, andδ in the following text represents the shrunken film thickness at the
wafer surface after PI drying and curing.
Table 1. Polyimide thickness over positive topography
Over Over stepflat area
Tbot Tbot − hm After spin-coating of PI
Shrunken PI thicknessδTbot δ(Tbot − hm + hs) due to high
temperature cure
δTbot δ(Tbot − hm + hs) + hm Total PI thickness
0 δhs + (1 − δ)hm Difference from flat area
can be seen in Fig. 2., the brand area at the bottom rep-
resents the wafer with its topographical features and theshaded area above the wafer of thickness Tbot represents
the thickness of the coating polyimide solution. Because
of the fact that polyimides are applied as liquids, the sur-face is kept flat by the surface tension (Fig. 2a). Shrinkagedue to the solvent loss or cure results in partial recovery ofthe underlying features (Fig. 2b) and thus the achievableDOP is controlled mainly by shrinkage [6]. From hereon only variations over a positive topography with heighthm are considered. The thickness of the solution over astep after spin-coating of polyimide is then Tbot − hm .
After polyimide coating has been dried and cured,shrinkage of the coating occurs, leaving a film with thick-ness δ(Tbot − hm + hs) + hm , which is a fraction of theoriginal thickness Tbot [3]. Here, also the contribution ofthe final polyimide step height is considered. Table 1 isshowing polyimide thickness variations over positive to-pography (ie, features which are above the surface of awafer).
To base our theoretical calculations on, we considerthat the thickness of the polyimide Tbot over the waferis much greater than the height of the features hm overwhich material will flow during cure.
When small steps are close enough together, the thick-ness over the steps will not be the same as over the plainareas. As reported in [3], the resultant thickness will bethe total volume of polyimide together with steps dividedby the area of the region. If this area has a repeating unit(eg , a chip), the polyimide thickness Tchip can be calcu-lated for a representative unit (Fig. 3a,b) as follows:
Tchip = δTbot + [δhs + (1 − δ)hm]AstepAchip
, (4)
where Astep is the area occupied by the steps in the chiparea.
Based on these assumptions, the pattern density Dcan be calculated as Astep/Achip , while the correspondingtwo areas can be expressed from the given length, widthand spacing of the steps.
Thus, referring to [3] with slight modifications, themodel of simple global planarization can be defined. Ingeneral, there are two cases which should be consideredfor the pattern density D :
(a) If we consider that area units (chips) are repeated intwo dimensions (see Fig. 3b), the global pattern densityfor the whole wafer will be
D =
∑Achip
Awafer, (5)
where Achip and Awafer are the total area of all chipsand the area of the whole wafer, respectively.
(b) On the other hand, if we consider only one-dimen-sional repeating of the steps (eg, in horizontal axis x)within a unit (chip), the formula for the local pattern den-sity for steps in that unit (see Fig. 3a) can be expressedas
D =
∑Astep
Achip=
w
w + x=
1
1 + xw
, (6)
-
88 L. Matay — R. Andok: PLANARIZATION OF MICROELECTRONIC STRUCTURES BY USING POLYIMIDES
(a)
(b)
w x
l
AstepAchip
A wafer
Fig. 3. a) Area of a chip consisting of long lines with length land width w . The spacing between these lines is marked as x .
b) Periodical distribution of the chips on a wafer.
Relative decrease of PI thickness
50 100 150 200 250 3000.7
0.8
0.9
1.0
Temperature regime I.
Temperature regime II.
Temperature (°C)
z
Fig. 4. Relative decrease of polyimide thickness versus temperaturedependence.
where w is the width of a feature and x is the spacingbetween two neighbouring features. As can be seen, in thiscase, the pattern density does not depend on the verticalsize of the features l , which is in good accordance with[4, 7], but is dependent only on the width of the linesunder the polyimide coating.
2 MATERIAL AND
EXPERIMENTAL PROCEDURES
In order to use polyimides as isolation materials itis important to know their planarization properties. Inour experiments a polyimide material of additional type
has been used which was synthesized in our institute andnamed as PPID. This polyimide solution has been dilutedin dimethylformamide (DMF) in accordance to the tech-nological requirements and three different solutions of thePPID polyimide have been used (ie, 30%, 35%, 40% and55% solutions).
Table 2. Testing modules for planarization
Testing module w [µm] l [µm] x [µm]I. 5 2 000 180II. 20 2 000 170III. 40 2 000 150IV. 100 4 000 300
For this purpose four different testing modules havebeen fabricated. Each module consisted of a set of five Allines with different widths and spacings between them.Table 2 is showing these modules for planarization; herew and l are the width and length of the lines, respectively,and x is the spacing between these lines.
The testing aluminium structures with thicknessesfrom 0.7 to 1.2 microns have been deposited by e-beamevaporation on cleaned silicon wafer, on which a 100 nmthick layer of SiO2 has been thermally grown. To mea-sure the profile variations of the evaporated aluminiumlines the Talystep measurement system by Rank-Taylor-Hobson with 2 nm resolution has been used. Then a so-lution of polyimide has been spin-coated using 3000 ÷5000 rpm for 30 seconds.
Experimentally, three types of thermal processes havebeen examined. All of the three processes had one thingin common, that is, that the wafers have been dried at140 ◦C for 20 minutes in an electrical oven with the aimof partial removal of the diluent.
The temperature regime I was characterized by divid-ing of the whole wafer in eleven parts which were in-serted into the electrical oven, where the temperaturegrew up continuously from 24 ◦C to 275 ◦C with the rateof 3 ◦C per minute. During the increasing temperature,these parts have been one after the other taken off fromthe oven and thickness of the polyimide film has beenmeasured. This process is shown in Fig. 4.
In the temperature regime II the wafer has again beendivided in eleven parts. Each part of the wafer with ap-plied polyimide film has been put in the oven with tem-perature interval set from 30 ◦C to 275 ◦C and taken offafter 12 minutes of curing. The thickness of polyimidefilm has been measured afterwards (Fig. 4).
As can be seen in Fig. 4, in the temperature regime I,where the thin polyimide film has a constant thickness,the temperature interval is between 25 ◦C and 185 ◦C.When this temperature has been exceeded the polyimidethickness decreases until it reaches 20 % less thickness (at272 ◦C) comparing to its initial value.
Comparing to the previous case, in the temperatureregime II, the thin polyimide film has a constant thicknesswithin the interval of 25 ◦C to 145 ◦C.
After reaching the temperature of 145 ◦C, the poly-imide thickness decreases until it reaches 19.42 % lessthickness (at 232 ◦C) comparing to the initial thicknessof the polyimide (before it was put in the electrical oven).The relative decrease of the polyimide film between the
-
Journal of ELECTRICAL ENGINEERING VOL. 53, NO. 3-4, 2002 89R
ela
tive
de
cre
ase
of
PI
thic
kne
ss
0 50 100 150 200 2500.70
0.75
0.80
0.85
0.90
0.95
1.00
55 percent
40 percent
30 percent
Time (min)
Relative decrease of PI thickness
Fig. 5. Relative decrease of polyimide thickness versus time of thethermal cure at constant temperature of the annealing for three
different concentrations of the PPID solution.
0.2 0.3 0.4 0.5 0.60.70
0.75
0.80
0.85
0.90
0.95
Under temperatureregime III.
Relative decrease of PI thickness
Concentration
Fig. 6. Relative decrease of the polyimide thickness versus corre-sponding concentration of the polyimide solution.
Degree of planarization
0 20 40 60 80 100
0.0
0.2
0.4
0.6
0.8
1.0
sample A
sample B
sample C
sample D
sample E
Line width (mm)
Fig. 7. DOP in dependence on the width of the planarized lines.
0.1 0.2 0.30
0.2
0.4
0.6
0.8
1.0
Pattern density
samples A-Esamples A-Csamples C,D
1 - DOP
Fig. 8. Planarization constant versus pattern density.
two temperature regimes is caused by different times ofcuring of polyimide films at the same temperature.
In the next experiment the influence of long-term con-stant temperature dependence on the relative thickness
change of polyimide films has been examined (tempera-ture regime III). Three silicon wafers have been coated
with three different concentrations (30 %, 40 % and55 %) of the PPID polyimide solution. All wafers have
been thermally cured at 150 ◦C for 30 minutes. The sam-ples prepared in this manner have been divided into eight
parts and annealed at 220 ◦C in nitrogen atmosphere. Allparts of the wafers have been gradually removed from the
oven and polyimide thickness has been measured. Thecorresponding data are plotted in Fig. 5.
Figure 6 shows experimental results of the relative
decrease of polyimide thickness in dependence on theconcentration of the polyimide solution. It can be seen
in Figs. 5 and 6 that the relative change of the polyimide
film thickness increases in accordance with the decreaseof the polyimide solution concentration.
3 RESULTS AND DISCUSSION
3.1. Results on local planarization
The dependence of DOP on the width of the planarizedlines for local planarization is illustrated in Fig. 7. As canbe seen in this figure, the dependence has a logarithmiccharacter. The dependence has been obtained from themeasured data and the resultant characteristic has beenfitted by Eq. 7 where α and β are fitting constants and wis the corresponding width of the planarized aluminiumlines.
As can be seen in Fig. 7, the curves labelled as A, B,C and D, E, respectively, show similar dependence. Ingeneral, for the same width of the planarization line, the
-
90 L. Matay — R. Andok: PLANARIZATION OF MICROELECTRONIC STRUCTURES BY USING POLYIMIDES
samples D and E were of a better degree of planarizationcomparing to samples A, B and C. This difference inthe DOP was caused by the different thickness of theplanarizing polyimide film (samples D and E had 1.56 µmand samples A, B and C had 1.36 µm of PI thickness, thehomogeneity of thickness used polyimide is below 1 %).
DOP = −α ln w + β , (7)
Theoretically, it can be said that the thicker the poly-imide film is, the better the degree of planarization canbe achieved.
3.2 Results on global planarization
Figure 8 shows the dependence of the planarizationconstant on the pattern density, calculated by Eq. 6. Itis evident that by increasing pattern density D also theplanarization constant will increase.
4 CONCLUSION
The planarization properties have been examined forthe new PPID polyimide material. Thin polyimide filmswere prepared by spin coating on to the surface of Si waferwith testing aluminum structures. The thermogravimet-ric analysis (TGA) has been used to evaluate the thermalprocesses in the PPID polyimide. Profilometer techniquehas been used to measure step height changes of isolatedlines with various dimensions. Theoretically, it can be saidthat a thicker polyimide film will lead to better degree ofplanarization. Also a model for local and global planariza-tion has been revised.
Acknowledgement
All experiments in present paper have been performedat the Institute of Informatics, Slovak Academy of Sci-ences in Bratislava through the VEGA projects No.2/7185/20 and No. 1/7620/20.
References
[1] LAI, J. H.—DOUGLAS, R. B.—DONOHOE, K. : Character-
ization and Processing of Polyimide Thin Films for Microelec-
tronics Application, Ind. Eng. Chem. Prod. Res. Dev. 25 No. 1(1986), 38.
[2] WONG, C.P. : Polymers for Electronic and Photonic Applica-
tions, Academic Press, Inc., Princeton, 1993.
[3] LAVERGNE, D.B.—HOFER, D.C. : Modeling Planarizationwith Polymers, SPIE 539, Advances in Resist Technology and
Processing II (1985), p. 115.
[4] ROTHMAN, L.B. : Properties of Thin Polyimide Films, J. Elec-trochem. Soc.: Solid-state Science and Technology 127 No. 10
(1980), 2216.
[5] CHAO, C.C.—WANG, W.V. : Planarization Enhancement of
Polyimides by Dynamic Curing and the Effect of MultipleCoating, Polyimides 2, Plenum Publishing Corporation (1984),
p. 783.
[6] SENTURIA, S.D. : Polyimides in Microelectronics, American
Chemical Society Symposium 346, Polymers for High Technol-ogy (1987), p. 428.
[7] WHITE, L.K. : Planarization Properties of Resist and Poly-
imide Coatings, J. Electrochem. Soc.: Solid-State Science and
Technology 130 No. 7 (1983), 1543.
[8] RAY, D.R.—RIDLEY, D.—MARIO, J.—SENTURIA, S.D. :
Polyimide Planarization in Integrated Circuits, Polyimides 2,
Plenum Publishing Corporation, 1984.
Received 12 October 2000
Ladislav Matay (Ing) was born in Piešt’any, Slovakia, in1950. He received the Ing degree in solid-state physics fromthe Faculty of Electrical Engineering of the Slovak Universityof Technology, Bratislava, in 1973. Since his graduation he hasworked at the Institute of Informatics of the Slovak Academyof Sciences in Bratislava. His research activities include ap-plication of thin polyimide films in microtechnology, mainlyin the field of microsystems technology and micromechanics(MEMS).
Róbert Andok (Ing) was born in Vel’ký Krt́ı̌s, Slovakia,in 1973. He received the Ing degree in solid-state physics fromthe Faculty of Electrical Engineering STU, Bratislava, in 1997.During the years 1997 and 2000 he was a PhD student at FEISTU in Bratislava, Department of Microelectronics; and sinceOctober 2000 he has worked at the Institute of Informatics ofthe Slovak Academy of Sciences in Bratislava, in the field ofmicrosystems technology.
E X P O R T - I M P O R T
of periodicals and of non-periodically
printed matters, books and CD - ROM s
Krupinská 4 PO BOX 152, 852 99 Bratislava 5,Slovak iatel.: ++ 421 2 638 39 472-3, fax.: ++ 421 2 63 839 485
e-mail: [email protected], http://www .slovart-gtg.sk
s.r.o.
GmbH
E X P O R T - I M P O R T
G.T.G.SLOVART s.r.o.GmbH
E X P O R T - I M P O R T
G.T.G.SLOVART