SMALL AREA MOLECULAR ANALYSIS AS APPLIED IN THE MICROELECTRONICS INDUSTRY

(1)

HAL Id: jpa-00223879

https://hal.archives-ouvertes.fr/jpa-00223879

Submitted on 1 Jan 1984

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are published 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.

SMALL AREA MOLECULAR ANALYSIS AS APPLIED IN THE MICROELECTRONICS

INDUSTRY

J. Ramsey

To cite this version:

J. Ramsey. SMALL AREA MOLECULAR ANALYSIS AS APPLIED IN THE MICRO- ELECTRONICS INDUSTRY. Journal de Physique Colloques, 1984, 45 (C2), pp.C2-881-C2-885.

�10.1051/jphyscol:19842202�. �jpa-00223879�

(2)

JOURNAL DE PHYSIQUE

Colloque C2, suppl6ment a u n02, Tome 45, f6vrier 1984 page C2-881

SMALL AREA MOLECULAR A N A L Y S I S AS A P P L I E D I N THE MICROELECTRONICS INDUSTRY

J. N. Ramsey

2/41C, IBM, HopeweZZ Junction, NY 22533, U.S.A.

RESUME

Les techniques de microanalyse Blgmentaire ont Bt6 appliquses avec succss comme aide au dzveloppement et _._. 2 la fabrication de la micro~lectronique depuis 24 ans.

Les analyses moldculaires seraient bien plus utiles, si elles pouvaient avoir une rSsolution spatiale comparable B la dimension d'un composant Qlementaire. Plusieurs de ces techniques, telles que la spectroscopie infrarouge ou Raman, la microscopic 5 lumiare polaris6e et la spectroscopie de massellaser seront discut6es et leur utilitz illustree dans le cas de contaminations organiques, occasionn6es par le contact avec la peau. Le Leybold-Heraeus LAMMA 1000, lorsque la puissance du laser sur 1'Bchantillon est gardse trSs faible etreproductible, donne des spectres frag- ment& aussi reconnaissables que ceux de la spectroscopie Blectronique de masse, dont les rzsultats sont largement publiss dans la littzrature.

ABSTRACT

Small area elemental analysis techniques have been successfully applied to the solution of microelectronics development and manufacture for 24 years. Molecular analysis would be far more useful if it could be done in device sized areas.

Several such techniques, infrared, Raman, polarized light microscopy and laserlmass spectroscopy will be discussed and their usefulness illustrated for organic contamination, including skin. The Leybold-Heraeus LAMMA 1000, when the laser power to the specimen is kept very low and reproducible, gives fragmentation spectra which are recognizable as those by electron impact mass spectroscopy, on which most/all library searches are based.

INTRODUCTION

as een 24 years to the month since Prof. David Wittr presented the first :Ep:icaFion of the electrop probe microanalyzer to the soyution of semiconductor/

microelectronics problems. The need for small area chemicgf,analysis has been strong in this field because of continuing miniaturization. This small area analyses, almost exclusive1

-g

elemental, have been essential for process development and roblem solvin

.

lemental information, however, is often not sufficient,

for exampfe, getting hysrogen, carbon, oxy en and nitrogen on a particle does not add measurably to its identification and efimination from a process.

Polarized light microscopy has lon been used in chemical and mineralogical studies, and was raised to a high Bevel of applicabilitg by Walter McCrone and associates, culminating in the extensive Particle Atlas

,

which included innumerable glasses, organics and polymers.

Small area molecular spectroscopies were obviously desirable, but there were problems of couplin a microscope and a spectrometer and retaining high e5ficiency ener y transfer. ~ E i s was solved for Raman by Delha e and Dhamelmcourt, which resufted in 1 micron la era1 resolution Raman capabifity in an instrument (the MOLE)

roduced b Jobln-Kvon.B This techni ue has been widely apptje12(several symposia gave been Xeld) includin~ the field of microelectronics. Small arealjnfra- red analysis was made possib e w ~ t h the introduction of the Nanometrics 20IR.

We have applieqqt&s extensively to process control problems in microelectronics manufacturlng. e use of the visible in Raman to see infrared transitions offers advantages. l l S r 9

Laser desorption/ionization and mass anal sis has been available inlihy9trans- mission mode for several years in the &L 500 by Leybold-Heraeus. This

techni ue has recent1 been extenqgd o the much more useful reflection mode as 1000 by ~ e ~ g o l d Heraeus '26 and the LIMA by Cambridge Mass Spectroscopy Co. It is evident that one of the major arameters to be controlled and measured is the laser power, to get reproducib?e fragmentation, hopefully recognizable to an organic mass spectroscopist and close to those from electron impact so that extensive literature/libraries can be utilized. If the laser power is too high and fragmentation extends to H, C, 0 and N, then little molecular information

is obtained, of course.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19842202

(3)

C2-882 JOURNAL DE PHYSIQUE

EXAMPLES OF-PROBLEM SOLVING BY MOLECULAR ANALYSIS a IC es on ecor m ur ace

h ~ e of ,r:' ieveiopEent ?lnEs"nai a problem, with -10um particulate contamination on a recording surface, which caused "dropped bits" on reading, and subsequent re- writing. In-situ small area Raman analysis (top two spectra of Figure 1) showed the material to be polypropylene (comparing it with the bottom spectrum). Poly- propylene, we were told, was not a possible contaminant because there was no goly- pro ylene in the materials of construction or packa ing. A further look at t e pacRage revealed that a paper-like material was use% to "clean" the surface on command. This material was "teased" apart and examined microscopically (insert Fi ure 2). There are small nodules on the fibers, seemin ly helping to hold the figers together. The fibers were shown to be cellulosic

Ey

conventional infra-red analysis and b polarized light techniques. The nodules were shown to be olypro- pylene by small area Raman (third spectrum, Figure 1) and by small area infra-red

(Figure 2). Was this indeed the source of the problem polypropylene articles?

We could not be sure, because while this was the only known source, tRere were no distinguishin elements unique to these pol propylenes (electron micro robe and LAMMA showed fi, but that is the .usual catafyst in poly ropylenes)

.

^{~ g e}"proof"

came when a different vendor's fllter paper (without nozules of any type) was used without subsequent problems.

What difficulties did we encounter with these analyses? The particles were embed- ding in the recordin medium and were difficult to remove, giving fragments which were difficult to coflect together on an infra-red trans arent substrate for small area infra-red: it failed. The fibers (a proximately lgum) were too small for small area infrared, and masses were too tElck: it failed. LAMMA fra mented the fibers and nodules too much: it failed (see later for LAMMA successes?.

11. Particulate Anal sis: Is it N lon or Skin?

Nylon an uman s In :re ver commoz cons 1 uen s of contamination in clean rooms (and not%-cleankrooms, tooy. The ~ylon~cEmes~from garments and bushingslgears, while humans are constantly e x f ~ l i a t i n e ~ s l f & : ~ g l a k e s and fragments. ave had great success with small areal&nfrared and small area RamanlEIQ

.

^However,

as R.M. Scott has pointed out

,

as Nylon and skin are both amides, their infrared spectra are uite similar, requiring full width at half maximum for differentiating.

Raman and po?arized microscopy were ex lored as improvements, with the latter being especially successful. The ver high gi-ref ringence of drawn Nylon (1.580 index of refraction axially, with 1.530 transverse) coupled with its distinctive hemi- spherical form with dimpled craters after melting (-265OC) allows unequivocal dlfferentiation from skin, with its low bi-refringence (-1.530 index of refraction) with no change of shape (but darkening in color) on being heated to -265'C). In addition the oils and salts in skin outgas at room temperature (and especially after -2.650~).

111. LAMMA 1000 Anal ses

As mentloned earllery the degree of fragmentation of a molecular ion controls the amount of useful information to lead to a molecular analysis, and I mentioned two exam les of excessive fragmentation. This precise control and measurement of very low Paser power a lied to the s ecimen is essential. Frank Anderson of IBM and Hans Heinen of ~&aeu~ h e d out such a method, and it will be discussed in their future paper. They have been kind enough to provide some of their early results on known organic materials, which enabled them to identify an unknown material within the scope of this present pa er, namely small area molecular analysis applied to micro-electronic device fabrication problems.

Thel,attempt was to get LAMMA fragmentation reproducible and to be recognizable to an or anlc" mass spectroscopist, by controlling and measuring the low laser power applie% to the specimen. Three sample materials, well characterized by Electron Impact/Mass Spectroscopy were chosen: Benzophenone, an aromatic ketone (Fig: 3), Rhodamine B, an indicator d e (Eig. 4) and Poly Alpha Methyl Styrene, an unzlppable polymer (Fig. 5). These

LA

successes" (defined as recognizable to EIIMS) gave confidence to try an unknown, a residue discovered after an etching and cleaning process in device manufacture. Figure 6 showed the material to be nitrobenzene sulfonic acid. (Note the overlap of the two mass range spectra, with 157 on both.) CONCLUSIONS

1 e ere are several small area molecular analysis techniques, the are usually

%mile&tar

,

with more than one required for a full analysls and soyution to process probPems: there is no universal technique.

ACKNOWLEDGEMENTS

am ln e e colleagues for sharing their data and ideas: F.W. Anderson -

;AMMA a ~ a ? ~ s ~ s F O D ~ ~ . Falcon

-

Small area olarized li ht analysis; H. Heinen (~e~bold-~erae;s)

-

LAMMA analysis- K.P. gadden

-

~ m a f l area and conventional inf ra- red analysis; C.D. Needham

-

~mall'area Raman analysis; R.M. Scott

-

Small area infra-red analysis; C.E. Wilson

-

Small area infra-red analysis.

(4)

RAMAN SPECTRA OF PARTICLES O N RECORDING SURFACE

Fig. 1. Raman S ectra of Particles, ~o$ul$s and Polypropylene Film

Fig. 2. Infra-red Spectrum of Nodules

BENZOPHENONE C13H100 NEGATIVE SPECTRUM FROM LAMMA- 100- 183 M+H

w-0

^M.182

EI/MS GIVES 182 (see text), 105 8 77 M - 105 PHENOL

I ^I

^I

1

183-Probably protonation of hydrogen due to high pressure

of material in laser beam.

7 7 PHENOL

I

1 , a

I,. L

s .a 3s 6a

---

⁷^-- 1.-

---

' A T O M I ~ MAS< UNIT< ' ' @ H ' F E I N E N ~ A N G ~ ~ ~ N - ~

Fig. 3. Negative S ectrum of Benzophenone by LAMMA 1000

The protonatlon

waul$

be ressure sensitive and may not occur at very low laser power, giving 182 as in

EP.

(5)

JOURNAL DE PHYSIQUE

NEGATIVE SPECTRUM FROM LAMMA 1000 RHODAMINE B Indicator dye

OH EI/MS GIVES 442, 397,

3 2 6 , 3 0 9 8 282

t. mm w ass >am a.a

A T O ~ I ~ ~ A S ~

'UNl;S

BK

HEINEN/&DERSON"--

Fig. 4. As indicated, all EI mass peaks are found in LAMMA, along with many extraneous peaks, which might confuse a novica mass spectroscopist, but ex erienced organic mass s ectrosco ists ran see" the pertinent peaks. This poPnts out the high level !ackgroun$ re uired in the early(?) stages of any analytical/characterization technique, ?or organics, especially.

a

POLY ALPHA METHYL STYRENE NEGATIVE SPECTRUM

Unzlppable polymer

Mrnonorner=118

118 MONOMER 105 (M-CH2)+H

!

I

EI/MS GIVES 118, 105,

Fig. 5

REFERENCES

1 ry, D.B., J.M. Axelrod and J.O. McCaldin, Proc. Semiconductor Conf. &ME iis!ok: August 31

-

Se tember 2, 1959.

*

^{2 Ramse} J.N., and P. Weinstein "?he Electron Mlcro robe" Editor R. McKlnle

.

^715-747.

*

3) Ehrist, J.G. and J.N. &sey: IEEE S ectrum 6 (k2~1?864;9!!~:

Pg

4) JCPDS, 1601 Park Lane, Swathmore, PA. 19081, ~ S A .

*

^{5 7}Cameron, D.P., F.W. Schneider,,and J.N. Ramsey, Proc. El. Mic. Soc. of Am 1969.

*

6) McCrone, W. and J. Delly, The Particle Atlas", Ann Arbor Sci. PU;~. Ann Arbor Mich., USA.

*

7) Delhaye and P. Dhamelin-

court J. Rama? S ectr. 3 (1975) 33. A 8) Jobin Yvon, 91160 Lon jumeau, FR.

*

9) ~'Actualite cgemique7; Avril 1980.

*

10) Proc. Microbeam ~ n a f ~ . Soc., 1978, 1981-3.

*

11) Needham, C.D., and J.N. Ramsey, Semiconductor International 4 (March 1981) 75.

*

12) Needham, C.D., Proc. SPIE (Soc. of Photo-Optical En ineers), Alexandria, VA. Meeting, A ril 1982.

*

13) Laring, D.J. and V.J. Coates, 5;oc.

SPIE 276 (1981) 249. lf Ramsex, J.N., and H.H. Hausdorff, Proc. 16th Meeting, M~crolZZm Analy Soc. (1981) 91 15) Scott R.M. and J.N. Ramsey, Proc. 17th Meetin ~icrobeam Analy. Soc. (i982) 239

*

¹⁶⁾Popek, K.M. and J.N. Ramsey, Proc.

SPIE ~fexandria, VA. Meeting A ril, 1981.

*

17) Andersen, M.E. and R.Z. Muggli, Anal Chem. 53 (Oct. 1981) i773:

*

18) Leybold-Heraeus, D-5000 Koln 51, FRG.

*

19 kufman, R., Proc. 17th Meetln Microbeam Analy. Soc. (1982) 341.

*

201 Hillencamp, F., Kaufman, R. anj'~. Wechsung, Proc. 18th Meetin Microbeam Analy.

Soc. (1983).

*

21) Cambrid e Mass S ectrometry, Ltd., Cambridge 6 ~ 4 4BH, England.

*

22) Evans, C.A., B.W. ~riffiths an$ T. Dingle, Proc. of 18th Meeting of Mlcro- beam Analy. Soc. (1983).

(6)

s e +

- - SPECTRA FROM LAMMA 1000

. -- -.

RESIDUE AFTER-

ETCHING

8 CLEANING 8 0 SO3 NITROBENZENE SULFONIC ACID

CgHgNOgS

M: 2 0 3 NO2

SO3, 157 M-NO2

1

F i g . 6

-

NEGATIVE SPECTRA

203 Parent (M)

C isotope peaks

~ -,v.T-?

,.-~.-