INSTITUTE OF PHYSICS PUBLISHING ...

[Pages:6]INSTITUTE OF PHYSICS PUBLISHING Nanotechnology 15 (2004) 435?440

NANOTECHNOLOGY PII: S0957-4484(04)72075-9

Non-contact nanoparticle removal with laser induced plasma pulses

Cetin Cetinkaya1 and M D Murthy Peri

Department of Mechanical and Aeronautical Engineering, Center for Advanced Materials Processing, Wallace H Coulter School of Engineering, Clarkson University, Potsdam, NY 13699-5725, USA

E-mail: cetin@clarkson.edu

Received 14 November 2003 Published 29 January 2004 Online at stacks.Nano/15/435 (DOI: 10.1088/0957-4484/15/5/006)

Abstract Nanoscale substrate cleanliness is an essential requirement in a variety of nanotechnology applications. The proposed particle removal technique based on pressure shock waves due to laser induced plasma is of interest in various nano/micro-manufacturing applications in which the minimum feature size is reducing rapidly. Any removal method adopted in a manufacturing process must be on the same shrinking feature reduction curve since, for device reliability, the minimum tolerable foreign particle size on a substrate depends on the minimum feature size on a nano/microsystem or device. In the current study, the transient pressure fields exerted on a surface with nanosecond pulse laser generated plasma shock waves are measured using a polyvinylidene difluoride film (PVDF)-based line transducer that was designed and tested for this particular measurement task. Using the pressure data, the corresponding diameters of latex particles that can be removed at these pressure levels are then calculated. It has been shown that latex particles as small as 60 nm in diameter can be removed from silicon surfaces using the nanosecond pulsed laser. Experimental removal data supporting these predictions are also included and discussed.

(Some figures in this article are in colour only in the electronic version)

1. Introduction

Nanoparticle removal is a critical requirement for various applications in nano-manufacturing, semiconductor devices and photonics/optics. As the feature size in nano/micromanufacturing continues to shrink, the particle removal requirements inevitably become more stringent. Industrial cleaning techniques currently in use include brush scrubbing, ultrasonic and megasonic cleaning, centrifugal spray cleaning, vapour phase cleaning, fluid jet cleaning and cryogenic cleaning. The brush scrubbing process is constrained by the size of the particle that has to be removed as the brush has to establish good mechanical contact with the particle [1]. In the megasonic and ultrasonic wet cleaning techniques, the fluid used should overcome the viscous effects and capillary force to generate high enough flow rates to remove the nanoparticles. For nanoparticles, excitation of the fluid to

1 Author to whom any correspondence should be addressed.

increase the efficiency of the process might lead to cavitations and subsequent damage [2]. The wet cleaning techniques such as centrifugal spray cleaning and fluid jet cleaning have to be followed by a heating process. During the heating process, the liquid may vaporize resulting in damage [1] and leaving stains of undesired chemicals [3]. The majority of these techniques are limited to surfaces without features (e.g. trenches and holes) [4] and substrate damage during their application is the major concern. In general, the size of particles that can be removed using these standard cleaning techniques is limited to 0.1 ?m. According to the 2002 International Technology Roadmap [5], by the year 2006, techniques that can remove standard polystyrene latex (PSL) particles with diameters less than 15 nm from silicon substrates will be required, and currently, according to the roadmap update, there is no known method for removing PSL particles smaller than diameter D = 40 nm. Therefore, the development of novel techniques that are dry, non-contact and non-destructive is needed for nanoparticle removal.

0957-4484/04/050435+06$30.00 ? 2004 IOP Publishing Ltd Printed in the UK

435

C Cetinkaya and M D Murthy Peri

Incident Beam

Shock Waves Focused Beam Convex Lens

Plasma Boundary

Blast Zone

d

Figure 1. A schematic diagram of the laser induced plasma (LIP) removal set-up (not to scale).

The use of short pulsed lasers for direct pulsed laser irradiation processes has been demonstrated and analysed in the past [6, 7]. The inertial forces generated due to the short pulse initiated rapid thermal expansion result in high surface acceleration [8], which leads to particle removal. The major concern with direct pulsed laser irradiation is the substrate damage due to strong localized stresses and thermal effects [9]. These stresses can result in substrate damage, change in surface roughness and cracking of the top layers. According to the linear thermoelastic material model, the approximate lower bounds for damage-free removal of silicon particles from silicon and copper substrates with a nanosecond pulsed laser are, respectively, 600 and 630 nm [9].

The laser induced plasma (LIP) process, developed as an alternative method to address the aforementioned shortcomings of the direct laser irradiation method, is a dry and non-contact particle removal technique. The use of LIP has been previously reported in [10, 11] for silica particles on silicon substrates and tungsten and copper particles on silicon wafers. The objective of the current study is threefold:

(i) to determine the effectiveness of the LIP process for nanoparticle removal;

(ii) to determine the minimum size for particles that can be removed with an available nanosecond pulsed laser; and

(iii) to compare the numerical and experimental results for PSL?silicon substrate systems.

As depicted in figure 1, in the LIP process an incident pulsed laser beam with 5 mm diameter is converged at the focal point of a convex lens.

This focusing leads to a rapid increase of temperature and energy density around the focal point leading to the dielectric breakdown of air and resulting in the formation of plasma [12]. In the case of a 7 ns laser pulse with 300 mJ pulse energy, it is reported [13] that the expansion and saturation of plasma takes place during the first few microseconds of breakdown and then the hot core of air surrounding the plasma emerges as an expanding spherical shock wavefront. In the LIP technique this shock wavefront pressure is directed to a surface with particles to break the adhesion bond between the particle and substrate system. The minimum size of particles that can be removed by LIP depends on the maximum pressure applied to the substrate, which is a function of the firing distance, the distance between the centre of plasma and the substrate [10]. The airflow induced by the shock wavefront could initiate a rolling and/or sliding of particles [14] if critical pressure magnitudes for the particles are achieved. Since the initiation of the pressure required for rolling is less than that for sliding, the dominant removal mechanism in the LIP process is assumed to be rolling.

z P ? As

Spherical Particle O a

D

2 r

D ? 2

0

Fa+ mg

Substrate

Figure 2. Forces acting on a spherical particle bound to the substrate during LIP removal.

The force of adhesion FA between a spherical particle and a flat substrate according to the Johnson?Kendall?Roberts (JKR) model is given by

FA

=

3 4

WA

D

(1)

where WA is the work of adhesion (Dupre?'s energy) between a spherical particle with diameter D and the substrate [15]. WA is the amount of energy required to separate the particle and the substrate from intimate contact established at the stable equilibrium configuration. The work of adhesion depends on the air humidity and temperature as well as the atomic structures and interface quality of the particle and the substrate materials (for details, see [16]). The work of adhesion between a PSL particle and silicon substrate under room conditions is given as WA = 2.02 ? 10-2 J m-2. The radius of contact between the spherical particle and the substrate surface a is determined, by considering elasticity, as

a = 3 WA D2 1/3

(2)

8K

where

K=4

(1 - 12) + (1 - 22)

-1

.

3 E1

E2

1, E1, 2 and E2 are, respectively, the Poisson's ratios and the Young's moduli of the particle material and the substrate material. Figure 2 depicts the force diagram of a spherical particle on a flat substrate. Applying moment balance at the point O gives the approximate relation for critical pressure required for removal of a particle in the rolling mode:

Pc

=

2a(FA + mg) As(D|cos | - 2a sin )

(3)

where As is the hemispherical area (effective area) normal to the applied LIP pressure, m the mass of the particle,

436

Laser Power Unit

Incident Beam

Trigger

Focused Beam

Plasma Boundary

d

PVDF Transducer

Oscilloscope

Charge Amplifier

Figure 3. A schematic diagram of the transient pressure measurement set-up in the laser induced plasma experiments.

Epoxy Mould

PVDF Active Sensing Area

Shim Backing 5 mm

0.75 mm

Non-contact nanoparticle removal with laser induced plasma pulses

figure 4. The rigid backing was moulded using fast cure epoxy (Epo-Kwick, Buehler). The mould was cased with an aluminium foil to minimize the electromagnetic interference between the laser beam and the film material since PVDF is pyroelectric. The equivalent capacitance of this transducer is measured as 4 nF and the sensitivity of the line transducer is determined as 0.0337 pC Pa-1. An inverting charge amplifier was designed and characterized using an arbitrary waveform generator specifically for this application. The bandwidth of the charge amplifier at -3 dB is determined as 0.19 (lower limit) and 41.6 MHz (higher limit). The PVDF line transducer was placed on a linear translation stage with a spatial resolution of 20?10 ?m. The critical distance d from the centre of plasma to the stage was adjusted by vertical sliding of the translation stage. The trigger line to the digitizing oscilloscope (Tektronix TDS 3052) was provided from the power unit of the Nd:YAG laser. The steps of the shooting process for the experiment are discussed below in detail.

Figure 4. The top view of the PVDF line transducer depicting the active sensing area of the transducer (not to scale).

g the acceleration due to gravity and the angle between the applied force and the plane parallel to the substrate surface. In the current application, the mg term in equation (3) can be neglected due to the magnitude of typically required acceleration levels for nanoparticle removal. The critical removal pressure for a 60 nm PSL particle on a silicon substrate is approximately 45.35 kPa when approaches 0 or [16]. To study the LIP transient pressure field on the surface of the substrate, a PVDF transducer integrated with a charge amplifier was employed.

2.1. Pulsed laser firing procedure

The PVDF line transducer was aligned with the plasma using a diode laser. A single pulse was fired at a distance of d = 1.3 mm and the PVDF line transducer connected to the charge amplifier generates a pulse proportional to the applied pressure since the central frequency of pressure excitation is low compared to the resonance frequency of the transducer. The transient waveform from the amplifier, triggered by the pulsed laser unit, was saved for processing. This measurement procedure is repeated for different firing distances and the recordings of the digitizing oscilloscope were saved.

2.2. Calibration for pressure waveform measurements

2. Experimental procedure

Initial calculations and removal experiments indicate that LIP is a viable technique for microparticle and nanoparticle removal [9, 11, 17]. To quantify the applicability of these predictions to nanoparticles, a set of experiments have been designed and conducted. A schematic diagram of the LIP transient pressure measurement set-up is presented in figure 3. The laser employed in the experiment is a Q-switched Nd:YAG (Quantel Brilliant series Q44) 1064 nm pulsed laser, operating at a fundamental wavelength of 1064 nm and with a pulse energy of 370 mJ. The pulse width is 5 ns, the repetition rate is 10 Hz and the beam diameter is 5 mm. The laser beam is converged using a 25 mm diameter, 100 mm focal length lens with a 1064 nm specific antireflective coating. A polyvinylidene difluoride (PVDF) line transducer was built using a 28 ?m thick, piezoelectric PVDF film (Measurement Specialties, Inc.). At low excitation frequencies compared to the lowest internal resonance frequency of the PVDF film, the voltage generated between the electrodes of the PVDF film is proportional to the magnitude of the pressure field exerted on a surface of the film. The PVDF film was bonded to a rigid plastic shim backing using cyanoacrylate adhesive. The effective dimensions of the rectangular sensing element are 0.75 mm ? 5 mm (which appears as a thin line in the top view) for good spatial resolution as shown in

Figure 5(a) shows the output waveform of the amplifier m(t) from the LIP measurement set-up. This output was transformed into the frequency domain, m( f ), as shown in figure 5(b) to study its frequency content. Figure 5(c)

represents the time?frequency analysis of the non-stationary waveform m(t) at a distance of d = 1.3 mm. Then, the pressure P( f ) at the PVDF line transducer is obtained from

P( f )

=

m( f ) T A( f )

(4)

where A( f ) is the amplifier gain function in frequency domain

and T the low frequency approximation of the transfer function

of the PVDF transducer, which is approximated for the frequency band of the current application as T = -g33t = 9.24 ? 10-6 V Pa-1 where g33 = -330 ? 10-3 V mPa-1 is the piezoelectric coefficient of the PVDF film for the axis of applied pressure and t = 28 ? 10-6 m is the thickness of the PVDF film. Finally, the response obtained, P( f ), is transformed back into the temporal domain, P(t), using an

inverse fast Fourier transform (IFFT) routine and the pressure waveform obtained, P(t), is presented in figure 5(d).

3. Experimental results

The LIP transient pressure response obtained at six firing distances, d = 1.3, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0 and 4.5 mm,

437

C Cetinkaya and M D Murthy Peri

m (V)

14 (a)

12 10 8 6 4 2 0 -2 -4 -6

0 2 4 6 8 10 12 14 16 18 20

Time (?sec)

(c)

log10(amplitude)

-4.6 (b)

-4.8

-5 m (f)

-5.2

-5.4

-5.6

-5.8

-6

-6.2

-6.4

0

0.2

0.4

0.6

0.8

1

1.2

Freq (MHz)

60

(d)

50

Laser Induced Plasma Pressure P(t)

40

30

P (KPa)

20

10

0

-1 0

-2 0

-3 0 0 2 4 6 8 10 12 14 16 18 20

Time (?sec)

Figure 5. (a) The voltage output m(t) of the amplifier from the LIP set-up at a firing distance d = 1.3 mm. (b) The frequency content of the pressure waveform m(t). (c) Time?frequency analysis of m(t) at a distance of d = 1.3 mm. (d) The transient pressure P(t) at a firing distance of d = 1.3 mm.

D (nm)

P (KPa)

50 40 30 20 10 0 -10 -20

2

d = 1.3mm d = 1.5mm

d = 2.0mm

d = 2.5mm d = 3.0mm d = 3.5mm d = 4.0mm d = 4.5mm

4

6

8

10

12

14

Time (?sec)

P (K Pa )

60

55

50

P 45

40

35

30

25

20

15

Damage Threshold

Particle Removal Limit

10

1

1. 5

2

2. 5

3

3.5

4

d (mm)

140 130 120 D 110 100 90 80 70 60 50 40 30 4. 5

Figure 6. LIP transient pressure responses on the surface at different firing distances from 1.3 to 4.5 mm.

are given in figure 6. The increase of the maximum pressure levels with the firing distance is clearly demonstrated. As the pressure level is increased, the minimum size of particle that can be removed by that level of pressure decreases as depicted in figure 7; the curve P denotes the LIP pressure peaks detected by the PVDF line transducers at different

Figure 7. The solid line shows the experimentally measured LIP pressure peaks (P) as a function of the firing distance (d). The dashed line corresponds to the minimum diameter (D) of the PSL particles that can be removed from a silicon substrate at these pressure levels according to the rolling criteria determined using the JKR particle?surface adhesion model.

firing distances. The pressure levels (P) generated in the LIP process and the corresponding PSL size (D) of particles that can be removed at different firing distances are also

438

Non-contact nanoparticle removal with laser induced plasma pulses

4. Conclusions and remarks

(a)

A set of experiments are conducted to study the nanoparticle

removal effectiveness of the laser induced plasma particle

removal technique for a 5 ns pulsed Nd:YAG laser. The

transient pressure fields generated in LIP are measured using

a PVDF line transducer, particularly designed and calibrated

for the current application. The LIP pressure fields at different

firing distances are measured and the pressure levels are related

to the minimum sizes of particles that can be removed. The

JKR adhesion model is used to determine the minimum PSL

nanoparticle that can be removed. These predictions are

compared to the experiments for removing 60 nm PSL particles

on silicon substrates. It is demonstrated and verified that the

LIP process can generate sufficient pressure fields for damage-

free removal of 60 nm PSL particles from silicon substrate. It

is noteworthy that the substrate damage reported in the current

(b)

study is predominantly due to the interactions between the

plasma and substrate material. The pressure levels measured in

this study are insufficient to cause mechanical surface damage

in silicon. Therefore, it is reasonable to conclude that, using

pulsed lasers with larger pulse energies and shorter pulse

widths, it would be possible to remove particles that are smaller

than 60 nm. Further work in this direction is under way.

Figure 8. (a) Before LIP and (b) after LIP SEM images at 5000? magnification. PSL particles with 60 nm diameters are identified in the before and after images. The dashed lines indicate the boundaries between the cleaning zones and location markings.

given in figure 7. The adhesion pressures required for PSL particle removal from silicon substrate were calculated using the JKR model. Thus, the curve D indicates the diameters of the PSL particles that can be removed corresponding to the LIP pressure generated at different firing distances. The curves P and D were extrapolated below a firing distance of d = 1.3 mm, since saturation of the charge amplifier occurred. In current experiments, it was observed that firing distances d 1.3 mm result in substrate damage for the type of laser used. Substrate damage in this particular experimental set-up is due to interactions between the plasma and the substrate. It is important to note that the first-damage mode is not mechanical. The damage threshold was marked at a firing distance of 1.3 mm in figure 7. At this firing distance, the size of particles that can be removed is approximated as 55 nm. As a result, the particle removal limit was marked at particle diameter D = 55 nm. This graph is in agreement with the published experimental results, which prove that 60 nm PSL particles can be removed at a firing distance of 1.4 mm as shown in figure 8 [17]. Thus, it is observed that the damagefree removal of 60 nm particles is possible with a 370 mJ pulsed laser at 1064 nm wavelength with 5 ns pulse duration in the LIP process.

Acknowledgments

The authors acknowledge the National Science Foundation (Nanoscale Exploratory Research Program, Award ID 0210242), the New York State Science and Technology Foundation and the Center for Advanced Materials Processing (CAMP) for financial support. Thanks also go to Ivin Varghese for providing the SEM images presented in figure 8.

References

[1] Wolf S and Tauber R N 2000 Silicon Processing for the VLSI Era-Process Technology 2nd edn, vol 1 (Sunset Beach, CA: Lattice Press)

[2] Mertens P W and Parton E 2002 Sub-100 nm technologies drive single-wafer wet cleaning Solid State Technol. 45 51

[3] Aoki H, Yamasaki S and Aoto N 1999 Cleaning for deep-submicron structures Electrochem. Soc. Proc. 99 102

[4] Vanderwood R 2002 MS Thesis Clarkson University, Potsdam, NY, USA

[5] The International Technology Roadmap for Semiconductors 2002 Update at

[6] Zapka W, Ziemlich W and Tam A C 1991 Efficient pulsed laser removal of 0.2 ?m sized particles from a solid surface Appl. Phys. Lett. 58 2217

[7] Lee J, Imen K and Allen S D 1991 A comparison of ns and ps steam laser cleaning of Si surfaces Appl. Phys. Lett. 58 203

[8] Cetinkaya C, Li C and Wu J 2000 Laser-based transient surface acceleration of thermoelastic layers J. Sound Vib. 231 195

[9] Lin J and Cetinkaya C 2003 Potential thermo-mechanical substrate damage in nanoparticle removal with pulsed lasers J. Adhes. Sci. Technol. 17 91

[10] Cetinkaya C, Vanderwood R and Rowell M 2002 Nanoparticle removal from flat substrates with pulsed-laser induced plasma and shock waves J. Adhes. Sci. Technol. 16 1201

[11] Lee J M and Watkins K G 2001 Removal of small particles on silicon wafer by laser-induced airborne plasma shock waves J. Appl. Phys. 89 6496

[12] Maker P D, Terhune R W and Savage C M 1962 Thermal effects and power enhancement in a He?Ne optical maser Alta Freq. 31 722

439

C Cetinkaya and M D Murthy Peri

[13] Sobral H, Villagra?n-Muniz M, Navarro-Gonza?lez R and Raga A C 2000 Temporal evolution of the shock wave and hot core air in laser induced plasma Appl. Phys. Lett. 77 3158

[14] Hooper T and Cetinkaya C 2003 Efficiency studies of particle removal with pulsed-laser induced plasma J. Adhes. Sci. Technol. 17 763

[15] Johnson K L, Kendall K and Roberts A D 1971 Surface energy and the contact of elastic solids Proc. R. Soc. 324 301

[16] Israelachvili J 2002 Intermolecular and Surface Forces 2nd edn (New York: Academic)

[17] Cetinkaya C and Varghese I 2004 Non-contact removal of 60 nm latex particles from silicon wafers with laser induced plasma J. Adhes. Sci. Technol. submitted

440

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download