Hoffman D.M., Singh B., Thomas J.H. (Eds). Handbook of Vacuum Science and Technology

Подождите немного. Документ загружается.

704

Chapter 5.4: Pulsed Laser Deposition

Fig.

Particle

free plume

Substrate

Target 1

Target 2

Shield

Micro particles

Cross plume configuration [31].

tides from the plume, but their practical value is limited because of their bulki-

ness and low transmission.

Another type of particle filter uses a second laser or a synchronous pulsed gas

jet perpendicular to the plume to "blow away" the slower particles [29]. However,

the efficiency of this approach is low.

Special Target-Substrate Geometry. Two special configurations have been re-

ported to reduce particle formation: a cross plume configuration and an off-axis

configuration.

a. Cross Plume Configuration

The cross plume configuration was first reported in 1975 [30] and later

modified [31] by Gapanov. In this approach, shown in Figure 5, two plumes

arrive at the same time and intersect in a region not far from the targets. Be-

cause of the high number density in the plumes, the coUisional frequency in-

side the intersecting zone is high. Kinematic restriction directs scattered

atomic and molecular species to move forward, while micron-size paiticles

from splashing continue their trajectories because of their heavy mass and

are therefore removed. This approach is applicable to all materials, but the

cross section of scattered plume is too small to be practical.

Off-Axis Substrate Configuration

The flow of ablation plume is hydrodynamic and has a radial component due

to the large number of collisions; however, the trajectories of heavy particles

do not change their courses. Therefore, in an off-axis configuration, the radi-

ally outdiffused species can deposit onto the substrate [9]. Films free of par-

ticles can be grown this way. This technique suffers from having a low depo-

5.4.5 Materials Survey 705

sition rate, and, most importantly, it negates some advantages offered by PLD.

In ZnO deposition using the off-axis configuration, particulate-free film with

a thickness uniformity of better than 5% over a 3-inch diameter area can be

grown [32]; however, its crystallinity is inferior to that of films grown in the

traditional face-to-face configuration. If a lattice-matching substrate is used,

then the dominant driving force for nucleation and coalescence is the lattice

energy between the film and the matching substrate rather than the plume im-

pingement energy. Under such circumstances, high-quality epitaxial films

can be grown using this configuration.

Liquid Target. The major source of splashing is the expulsion of molten

droplets from the surface melt by the recoil pressure of the departing plume. Ex-

pulsion takes place at the liquid-solid interface near the edge of the laser spot.

Therefore, one way to eliminate this effect is to make the area of the melt much

greater than the laser beam spot size. By the time the oscillatory motion of the

melt surface reaches the liquid-solid interface, its amplitude has been sufficiently

damped to prevent any splashing. The principle was first verified by Sankur [33] in

a brilliant attempt to grow Ge films by PLD using a molten Ge source. Particulate-

free films could be grown even under extremely high-energy fluence. Recently

Xiao [34] reported the growth of diamond films by PLD, using diffusion pump oil

as a liquid target source. The approach is unique and successful. Although the use

of a liquid target is scientifically sound and technically simple, it suffers from the

major drawback of material availability. As a matter of fact, very few materials

satisfy the criterion of having low vapor pressure at the melting temperature.

This obstacle was recently overcome by Witanachchi [35] with a dual-laser ap-

proach in which a TEA-CO2 laser was used in conjunction with a KrF laser. The

pulses are synchronized such that the target surface is melted by the CO2 laser

just prior to the arrival of a KrF laser pulse for ablation. The CO2 laser is defo-

cused, covering an area much larger the focused KrF laser spot. This approach is

extremely promising. It can be applied to all materials of interest and has an

added advantage in achieving higher plume plasma temperature and better unifor-

mity over a large area.

5.4.5

MATERIALS SURVEY

PLD demonstrates its versatility in a wide material selection. The first list of

PLD-grown materials was compiled in 1988 in a review article by Cheung and

Sankur [1]. Since then, owing to the growing popularity, the list has grown con-

siderably [2]. Here, we will only highlight some key examples representing

dif-

ferent families of materials and point out some trends.

706 Chapter 5.4: Pulsed Laser Deposition

5.4.5.1 Metals

Most work on metal films grown by PLD was reported in the 1960s and 1970s

[36].

Since PLD does not present any obvious advantage in this area over com-

petitive techniques, studies on metal film growth by PLD have slowed. Only in

areas such as thin-film coating of magnetic metal alloys [37] and other non-

equilibrium alloys [38] where stoichiometry plays an important role, does PLD

have an advantage over the conventional techniques.

5.4.5.2 Semiconductors

There are some advantages for using PLD to grow semiconductor films, but there

is also one major disadvantage: the presence of micron particles from splashing.

The only semiconductors that do not splash are the II-VI compounds, because

they lack a liquid state. By no coincidence, PLD growth of II-VI semiconductors

has been very successful. This work can be exemplified by the series of studies on

Hgi-xCxdTe, CdTe and their superlattices [39] and ZnSe and ZnSe films for so-

lar cell application [40]. More recently, Lowndes has demonstrated the growth of

p-type ZnSe film by in situ nitrogen doping [41], opening up novel applications.

Overall, the current trend suggests that PLD-grown semiconductor layers will re-

main in these niches and must face stiff competition from well-developed tech-

nologies such as MBE and MOCVD.

5.4.5.3 Diamondlike Carbon

Diamondlike carbon film is a composite network of an SP^ (graphitic) and SP^

(diamond) carbon. Although other techniques such as C VD and ion beam deposi-

tion have been used to grow this material, the qualities of thermal conductivity,

microhardness and field emission efficiency are still inferior to PLD DLC films

[42].

In addition, PLD has been used to grow CN^ films [15] exhibiting the high-

est nitrogen content.

5.4.5.4 Perovskites

Perovskite is a family of oxides with a general formula of ABO3 where A and B

represent cations. Deposition of perovskites has been the traditional stronghold

of PLD and will continue to be so. PLD owes its rapid growth in the late 1980s

to the successful demonstration of YBCO films. In the early 1990s, the YBCO

work opened the door for ferroelectric films studies [21]. Recent activity in co-

5.4.5 Materials Survey 707

lossal magneto resistance (CMR) materials is a further extension of this research

area [43].

The structural properties of perovskite materials and the advantages of PLD

match up extremely well. In fact, PLD has served as a trail blazer in the new ma-

terial frontier involving perovskite thin films. However, we should re-emphasize

that PLD growth of perovskite thin film is not a large-area process. Scaleup to

larger substrates is underway in a number of laboratories.

5.4.5.5 Simple Oxides

Oxide films are the most logical material of choice for PLD, especially "simple

oxides" referring to those containing only one type of cation. In this area, to be

successful PLD must compete against E-beam deposition and sputtering, in par-

ticular RF magnetron sputtering. PLD has met such challenges and demonstrated

a definitive advantage. Simple oxide film growth by PLD has flourished and is

firmly established in this niche. The highest-impact work in this area is the series

on heteroepitaxy of oxide/semiconductor structures by Fork [44]. In particular,

the epitaxy of yttria-stablized zirconia (YSZ) on Si [45] and the epitaxy of MgO

on GaAs [46], two materials of technological relevance, are important examples.

5.4.5.6 Nitrides

The main challenge in growing nitride films is to overcome the barrier of produc-

ing high concentrations of active nitrogen. PLD has proven successful in doing

so.

The first successful demonstration in this area is the growth of TiN [47], fol-

lowed by poly crystalline AIN [19], BN [48], and the growth of GaN in high vac-

uum [49]. The main rival in this area is MOCVD, which has been very actively

used for nitride growth in the recent year.

5.4.5.7 Organic Materials

Clearly, the growth of organic materials must involve a nonthermal process to

avoid dissociation. For this reason, UV PLD is a logical candidate. The ablation

mechanism of organic polymers is quite different from that of other materials.

The mechanism is mainly photolytic, involving "zipping" polymer chains into

small units and repolymerizing them on the substrate [50]. Examples of materials

deposited include polyethylene, polycarbonate, polyimide, polymethylmethacry-

late (PMMA) [51], polyphenylene sulfide [52], and even Cu-phthalocyanin films

[53].

708 Chapter 5.4: Pulsed Laser Deposition

5.4.6

FUTURE OUTLOOK

PLD has made giant strides in the last 10 years and has undergone a metamor-

phosis from an obscure and curious laboratory innovation to a powerful thin-film

deposition technique that is widely accepted by the technical community. There

have been numerous articles, books, workshops, and symposia on this subject. In

terms of material selection, it has stepped out from the early success with perov-

skites and broadened the horizon to other systems. The availability of commercial

PLD systems is another sign of maturity. Along a parallel path, in situ diagnosis

of the laser ablation process has evolved rapidly to provide insightful information

relevant to process control. While UV excimer lasers used to be the weak link of

this process, because they lack long-term stability, they have now been improved

to a point comparable to other thin-film deposition hardware. However, despite all

this progress, PLD still remains a research tool. To join the mainstream of tech-

nology, PLD must address and solve the issues concerning large-area scaleup and

morphology improvement. It must also identify niches where the performance of

this technique is unique and superior to other techniques.

REFERENCES

Jeffrey T. Cheung and Haluk Sankur, CRC Critical Review on Solid State Material Science, 15

(1988)63-109.

Pulsed Laser Deposition of Thin Films, edited by Douglas B. Chrisey and Graham K. Hubler

(Wiley, New York, 1994).

A. F. Turner and H. M. Smith, Appl. Optics, 4 (1965) 147-148.

D. Dijkamp, T. Venkatesan, X. D. Wu, S. A. Shaheen, J. Jisrawi, Y. H. Min-Lee, W. L. McLean,

and M. Croft, Appl. Pliys. Lett., 51 (1987)

619-621.

V. S. Ban and D. A. Kramer, J. Material Science, 5 (1970) 978.

6. H. Schwartz and H. A. Tourtellotte, J.

Vac.

Sci. Technoi, 6 (1969)

3763.

P. Meir-Komor, Niicl. Instrum. Methods, 167 (1979) 73.

8. Yu-A Bykovskii, V. M. Boyakov, V. T. Galochkin, A. S. Mochanov, I. N. Nokolaev, and A. N.

Orevskii, Sov. Phys. Tech. Phys., 23 (1978) 578.

9. B. Holzapfel, B. Roas, L Schulz, P. Bauer, and G. Saemann-Ischenko, Appl. Phys. Lett., 61

(1992)3178.

10.

R. J. Kennedy, Thin Solid Films, 214 (1992) 223.

11.

Mamoru Iwabuchi, Kazuya Kinoshita, Hiroshige Ishibashi, and Takeshi Kobayashi, Jpn. J.

Appl.

Phys., 33 (1994) L610.

12.

K. Maruyama, S. Endo, G. Sasaki, K. Kamata, J. Nishino, and K. Kuchitsu, J. Mat. Sci. Lett., 11

(1992) 1588.

13.

H. S. Kwok et al., Appl. Phys. Lett., 59 (1991)

3643.

14.

Seth Jashree, R. Padiyath, T H. Rasmussen, and S. V. Babu, Appl. Phys. Lett., 63 (1993) 473.

References 709

15.

C. Niu, Y. Z. Lu, and C. M. Lieber, Science, 261 (1993) 334.

16.

J. T. Cheung and J. Madden, J.

Vac.

Sci.

Techno!.,

B5 (1987) 705.

17.

T. Kawaii, Digest,

lEEE/LEOS

J 996

Summer Topical

Meeting,

Advanced Applications of Lasers

in Materials Processing, (IEEE, Keystone, CO, 1996), pp. 3-4.

18.

See David B. Geohegan, Diagnostics and characteristics of pulsed laser deposition laser plasma,

in Reference 2, chap. 5.

19.

G. M. Norton, P. G. Kotula, and C. B. Carter, Appl. Pliys. Lett., 70 (1991) 2871.

20.

X. X. Xi, C. Doughty, A. Walkenhorst, et al., Phys. Rev. Lett., 68 (1992) 1240.

21.

R. Ramesh et al., Appl. Phys. Lett., 57 (1990) 1505.

22.

J. A. Greer, J.

Vac.

Sci. Technoi, AlO (1992) 1821.

23.

Roger Kelly and J. Rotheberg,

Niicl.

Inst, and Methods, B7/8 (1985) 755.

24.

J. F. Ready, Appl. Phys. Lett., 3 (1963) 11.

25.

J. T. Cheung, Appl. Phys. Lett., 43 (1983) 255.

26.

D. Lubben, S. A. Barnett, K. Suzuki, S. Gobatikin, and J. E. Greene, J. Vac. Sci. Technoi, B3

(1985)968.

27.

A. B. Beavitt, Thin Solid Films, 1 (1967) 69.

28.

H. Dupendant, Appl.

Surf.

Sci., 43 (1989) 369.

29.

Kouichi Murakami, Laser Ablation of

Electronic

Materials —Basic

Mechanisms

and Applica-

tions, edited by E. Fogarassy and S. Lazare (Elsevier Science, 1992) pp. 125-140.

30.

S. V. Gaponov, A. A. Gudkov, and A. A. Fraerman, Sov. Phys. Tech. Phys., 11 (1982) 1130.

31.

M. D. Strikovsky, E. B. Klyuenkov, S. V. Gapanov, J. Schubert, and C. A. Copetti, Appl. Phys.

L^rr.,

63 (1993) 1146.

32.

J. T. Cheung (unpublished results).

33.

H. Sankur, W. J. Gunning, J. DeNatale, and J. FlintofF, Appl. Phys. Lett., 65 (1989) 2475.

34.

R. R Xiao, Appl. Phys. Lett., 69 (1995) 1022.

35.

Sarath Witanachchi and Pritish Mukherhee, in Advanced Laser Processing of

Materials —

Fun-

damentals and Applications, MRS Symposium Proceeding Vol. 397, edited by R. Singh, D. P.

Norton, L. D. Laude, J. Narajan, and J. T. Cheung (Material Research Society, Pittsburgh, 1996).

36.

V. M. Boyakov, V. M. Epikhin, B. A. Kalin, M. K. Makhatov, I. N. Nikolaev, and G. N. Shiskin,

Sov. J.

Quantum

Electron., 8 (1978) 902.

37.

K. Hayashi, M. Hayakawa, W. Ishikawa, Y. Ochiai, H. Matsuda, Y. Iwasake, and K. Aso,

J. Appl. Phys., 61 (1987) 3514.

38.

F. R. DeBoer, R. Boom,

C. M. Mattens, A. R. Miedema, and A. D. Niessen, Cohesion in Met-

als (North-Holland, Amsterdam, 1988).

39.

J. T. Cheung, G. Niizawa, J. Moyle, N. P Ong, B. Paine, and T. Vreeland, Jr., J.

Vac.

Sci. Tec/i-

;7o/.,A4 (1986) 2086.

40.

A. Aydinli, A. Compaan, G. Contreras-Puente, and A. Mason, Solid State Comm., 80 (1992a)

465.

41.

C. M. Rouleau, D. H. Lowndes, J. W. Macamy, J. D. Budai, D. B. Poker, D. B. Geohagen, A. A.

Puretzky, and Shen Zhu, Appl. Phys. Lett., 67 (1995) 2545.

42.

For a review, see A. Rengan and J. Narajan, Laser Ablation of

Electronic

Materials —Basic

Mechanisms and Applications, edited by E. Fogarassy and S. Lazare (Elsevier Science, Amster-

dam 1992), pp. 363-376.

43.

J. H. Hao, X. T. Zeng, and H. K. Wong, J. Appl. Phys., 79 (1996) 1810.

44.

David Fork, Epitaxial oxides on semiconductors, in Reference 2, chap. 6.

45.

D. K. Fork, D. Fenner, G. A. N. Connell, J. M. Phillips, and T. H. Geballe, Appl. Phys. Lett., 57

(1992)1137.

46.

W. Presseit et al., Appl. Phys. Lett., 46 (1992) 1499.

47.

J. Narajan, Appl. Phys. Utt., 61 (1992) 1290.

48.

G. L. Doll, J. A. Sell, C. A. Taylor II, and R. Clarke, Phys. Rev., B43 (1991) 6816.

710 Chapter 5.4: Pulsed Laser Deposition

49.

R. F. Xiao, H. B. Liao, N. Cue, X. W. Sun, and H. S. Kwok, J. AppL Phys., 80 (1996) 4226L.

50.

R. Srinivisan, B. Braren, and K. G. Casey, J. AppL Phys., 68 (1990) 1842.

51.

S. G. Hansen and T. E. Robitaille, J. AppL Phys., 84 (1988b) 2122 and AppL Phys. Lett., 52

(1988)81.

52.

S. Kale, M. Swaminathan, and S. B. Ogale, Thin Solid Films, 206 (1991) 161.

53.

F. Kannari, N. Matsumoto, E. Ina, and T. Fujii, Digest lEEE/LEOS 1996 Summer Topical Meet-

ings {Advanced Applications of Lasers in Material Processing) (IEEE, Keystone, CO, 1996),

63.

CHAPTER

5.5

Plasma-Enhanced Chemical

Vapor Deposition

Frank Jansen

BOC

Coating Technology

5.5.1

INTRODUCTION

Some thin-film deposition processes use molecular gases and vapors, rather than

solid evaporants or sputter targets, as their source materials. These processes are

generally referred to as "chemical vapor" deposition processes, implying that the

starting gases are "chemically activated" before deposition. In plasma-enhanced

chemical vapor deposition (PECVD), chemical activation is achieved by supply-

ing electrical power to a gas at reduced pressure, typically between 10 mtorr and

10 torn At these pressures, the application of a sufficiently high voltage creates a

visible glow, called glow discharge plasma. The plasma consists of about equal

concentrations of ions and electrons. The visible glow is caused by charge re-

combination processes and the relaxation of electronically excited atoms and

molecules. The electrical power is coupled into the gas through the mediation of

plasma electrons. The energetic electrons in the plasma ionize the gas, if only to

a minor extent—about one part per million. A much larger fraction of the gas,

about

percent, is chemically activated by the electrons. The increased chemical

activity of the gas results primarily from dissociation of the molecules into smaller

species, called radicals

[1,2].

Radicals are chemically unsaturated and therefore

^^^•^^ All rights of reproduction in any form reserved.

711

712

Fig.

Chapter 5.5: Plasma-Enhanced Chemical Vapor Deposition

electron

gas molecule

radical (

radical ^

chemical

activation

transport

The basic elements of the PECVD process: (a) chemical activation by electron impact, e.g.,

molecular dissociation; (b) transport ol the activated species to the film surface, e.g., by diffu-

sion;

and (c) bonding to the film surface.

capable of chemical reactions at high rates; they are the species that react at a sur-

face and contribute to film formation.

One can picture the PECVD process as a sequence of three distinct processes:

the chemical activation of

gas molecule through electron impact dissociation, the

transport of the radical species to the substrate, and the chemical reaction at the

film surface (Figure 1). We will discuss each of these processes in some detail.

The relatively light electrons present in the plasma are easily accelerated by the

electric field. The same electrical force acts on the ions but they are much heavier

and accelerate slower. Electrons don't lose much energy in most collisions with

the background gas. Due to the large mass difference between the electrons and

the gas molecules, low-energy electrons collide elastically with the gas mole-

cules,

like marbles bouncing off a billiard ball. A plasma electron can therefore

accumulate more and more energy in the electric field. Once an electron has gained

enough energy, an inelastic collision may take place. Like a bullet shot into a bil-

liard ball, the electronic bullet loses energy and the molecular billiard ball in-

creases its energy and often dissociates. In such an inelastic collision, the electron

loses several electron volts (eV's) of energy to the gas molecule—Figure 2(a).

This energy can cause a gas molecule to rotate and vibrate; it can also electroni-

cally excite the gas or cause it to dissociate or ionize. Many of these inelastic in-

teractions increase the "chemical activity" of the gas molecules by several orders

5.5.1 Introduction

713

Fig.

electron energy

ccelerates

bollides

accelerates

/collides

/accelerates

/collides

/accelerates

/collides

/accelerates

/collides

/accelerates

/collides

(accelerates

I collides

f accelerates

collides

inelastic

collision

etc.

Dissociation

Ionization

Vibration

Rotation

Electronic Transition

(a)

dc: secondary electron

rf: plasma electron

^w: "free* electron

time

N(E)dE

(b)

electron energy

dissociation energies

ionization energies

15 eV

(a) The energy vs. time for a single electron undergoing numerous elastic collisions and an in-

elastic collision; (b) A typical energy distribution of the plasma electrons. N{E)dE is the num-

ber of electrons with energy between E and E + dE.

of magnitude. The ensemble

plasma electrons

characterized

by a

(nearly)

Maxwellian electron energy distribution

[3]

— Figure 2(b).

The mean electron temperature

of a

typical glow discharge

equivalent

to an

energy of between

and 3

eV.

This energy must be compared to the energy needed