Arthur Sherman - Chemical Vapor Deposition for Microelectronics (779637), страница 31
Текст из файла (страница 31)
At the same time, silane flows into the chamber adjacent to butProduction CVD Reactor Systems173not in the discharge chambers. In the latter, the N2 dissociates, and becauseof its long recombination time, N atoms are available to react with the silaneon the wafer surface. Because of this pre-ionization and dissociation of N2 ,it is not necessary to heat the wafer to promote the reaction at reasonablerates.
In an arrangement such as this, there will be little ion bombardmentof the wafer during deposition, If such bombardment were desired (i.e., enhance compressive stresses), a second electrode can be powered, as shown,to create a plasma around the substrates.As noted earlier, this is the only system on the market that can depositgood quality silicon nitride films at room temperature, As low-temperatureprocessing becomes more valuable, this approach will attract more and moreattention..'."Ot,.lMlt' ..
•":..-..-="il"!•---,,'_"'to:r.\;\:-,.(,11Figure 26: Low-temperature PECVD reactor system-Ionic Systems, Inc.174Chemical Vapor Deposition for MicroelectronicsProcesschamberAF leeatnroughSubstratesAtomller gasballie-elecirode.............- + - - - Atomizer cavityAtomizer cav11ygas mletThrOnleval~PumpoulbalfleMalOchambefhoi electrodesaeet1Ground screenFigure 27: Reaction chamber schematic-Ionic Systems, Inc.REFERENCES1. Benzing, W.O., Rosier, R.S., and East, R.W., A production reactor forcontinuous deposition of silicon dioxide.
Solid State Technol. 16:37(1973).2. Winkle, L.W., and Nelson, C.W., Improved atmospheric-pressure chemicalvapor-deposition system for depositing silica and phosphosilicate glassthin films. Solid State Technol. 24(10): 123 (1981).3. Sherman, A., Design of plasma processing equipment. To be published.7Film Evaluation Techniques7.1 INTRODUCTIONThroughout all the preceding chapters, we have discussed thin films thatcan be created by chemical vapor deposition in terms of their physical andchemical attributes.
However, we did not explain how we secured the necessary physical or chemical information. For example, we discussed film deposition rates many times, but did not explain how we knew the film thicknessafter a specified amount of time. Similarly, when we spoke of the stoichiometryof deposited composite films, we did not indicate how we determined theirchemical composition.In the present chapter, we will attempt to correct this oversight. The firsthalf of the chapter will review the many techniques whereby we measure thephysical nature of the film we have deposited. The second half wilt cover thechemical composition of the film, both in bulk (average over film thickness)as well as how it varies through the film thickness.Since the measurement techniques for thin films from several micronsdown to several hundred Angstroms thick are quite sophisticated, it was feltthat their detailed description would be better left to a separate chapter.
Inthis way, they can be dealt with in some detail without interfering with thestudy of the various CVD techniques.7.2 PHYSICAL MEASUREMENTSIn this section, we will discuss those techniques one uses to evaluate thephysical characteristics of the thin films we can deposit. We specifically deferquestions as to the chemical nature of the film.7.2.1 Th icknessThe measurement of film thickness can be a fairly simple measurement175176Chemical Vapor Deposition for Microelectronicsor it can be quite complex, depending on the nature of the film. The mostdirect technique is the measurement of the step height when a portion of thedeposited film is etched away.
This is done by electronically tracking the position of a mechanical stylus as it is traversed across the step. Such a surfaceprofilometer is illustrated in Figure 1.A typical surface profile is shown on the video display. Vertical resolution of 5 A and horizontal resolution of 400 A is claimed. As long as the deposited film can be etched off the substrate without etching the substrate,this technique can be used for any thin film. Its primary utility is for R&Dstudies, as it is clearly not a production technique. The only film for which itis not suited is an epi silicon film on a single-crystal silicon substrate. A technique for measuring the thickness of these films will be described in the section on Infrared Spectroscopy.Figure 1: Computerized surface profilometer, Alpha-Step 200 Tencor Instruments.Film Evaluation Techniques177As long as the film is not reflective (i.e., specular aluminum) and is deposited on a reflective substrate (i.e., Si0 2 on silicon), optical techniques areavailable.
It was recognized early that the color of a thin film could be correlated to its thickness. Although not very precise, such information is veryuseful for quick evaluation in the laboratory. For example, silicon dioxidefilms on silicon substrates can be evaluated with the data of Table 1. In fact,one of the more useful aspects of this technique is that one can make rapidjudgements as to film uniformity.Going beyond this simple qualitative technique, the thickness of filmscan be measured by a polarizing spectrometer or "ellipsometer." This is aninstrument whose operation is based on the fact that elliptically polarizedlight changes its polarization upon reflection from a thin transparent filmon a reflecting substrate. The ellipsometer creates an elliptically polarizedmonochromatic light beam, and then evaluates the light beam on reflectionfrom a thin film.
The essential ingredients from an ellipsometer are shown inFigure 2. 2 A monochromatic beam of light (today most often from a laser)passes into a polarizer where it becomes plane polarized. It then passes througha compensator which converts it into an elliptically polarized light beam. Afterreflection from the substrate/thin film, it passes through an analyzer. If it hadbeen converted back to plane polarized when it had been reflected, then itwould be possible to rotate the analyzer to find a true minimum intensity.The technique then is to adjust the polarizer until the reflected light is planepolarized.
The analyzer is rotated to determine the position correspondingto a minimum in light intensity. This information, along with a theoreticalmodel of the optical process almost 100 years old, permits a calculation ofthe film thickness.With the advent of modern computing capabilities, ellipsometers havebeen automated and have proven useful in production settings. Originally,this technique was found most useful for the evaluation of dielectric filmsdeposited on silicon substrates. Today, more sophisticated instruments suchas the one shown in Figure 3 can be used to measure a wide variety of thinfilms on many different substrates.
Even metal films can be measured if theyare less than 500 A thick.Finally, we should note that in addition to 'film thickness, the index ofrefraction of the film can be determined and used to obtain chemical information about the film. This aspect will be discussed in Section 7.3.1.Another instrument widely used to measure film thickness is a spectrophotometer that operates over the visible light (4800 to 8000 A) wavelengthrange. This instrument essentially quantifies the qualitative evaluation of filmcolor mentioned earlier.
A commercial instrument operating on this principleis shown in Figure 4.Light reflected from the thin film is passed through the optical microscope onto a dispersive grating. The grating is then mechanically rotated sothat the light spectrum is passed over a thin slit. The intensity of light passingthrough the slit is measured by a photointensity meter and recorded by theCOrTlputer. In this way, the most intense frequency (color) is determined. Thisinformation, plus knowledge of the index of refraction, allows the film thickness to be determined.Table 1: Si0 2 Thickness vs.
Color!FILMORDERTHICKNESS(MICROMETERS) (5450 A)COLOR ANDCOMMENTS0.0500.075tanbrown0.1000.1250.1500.175dark violet to red violetroyal bluelight blue to metallic bluemetallic to very light yellowgreenIlight gold or yellow-slightlymetallicgold with slight yellow orangeorange to melonred violet0.2000.2250.2500.275FILMTHICKNESS(MICROMETERS)0.5020.5200.5400.5600.5740.600.630.680.720.770.800.820.850.860.870.890.390yellow greengreen yellowyellow0.920.950.970.990.4120.4260.4430.4650.4760.4800.493light orangecarnation pinkviolet redred violetvioletblue violetblue1.001.021.051.061.070.3650.375IIIII0.585blue to violet blueblueblue to blue greenlight greengreen to yellow green0.3000.3100.3250.3450.350ORDER(5450 A)IVCOLOR ANDCOMMENTS-""'-JCOLOR ANDCOMMENTS00()::::::rblue greengreen (broad)yellow gret:ngreen yellowyellow to"yellowish,,*light orange or yellowto pink borderlinecarnation pinkviolet red"bluish"**1.101.111.121.181.19blue green to green (quite broad)"yellowish"1.32orange (rather broad for orange)salmondull, light red violetvioletblue violetblueVFIL.\{ORDERTHICKNESS(MICROMETERS) (5450 A)VIviolet redcarnation pink tosalmonorange"yellowish"1.211.241.251.281.401.451.461.501.54greenyellow greengreenvioletred violetVIIVIIIsky blue to greenblueorangevioletblue violetbluedull yellow greenblue greendull yellow greenyellow to "yellowish"orangecarnation pinkviolet redred violetvioletblue violet*Not yellow, but is in the position where yellow is to be expected; at times it appears to be light creamy grey or metallic.*"Not blue but borderline between violet and blue green; it appears more like a mixture between violet red and blue green and overall looks greyish.NOTE: Above chart may also be used for Vapox, Silox, and other deposited oxide films.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
