In situ laser processing in a scanning electron microscope

7
In situ laser processing in a scanning electron microscope Nicholas A. Roberts, Gregory A. Magel, Cheryl D. Hartfield, Thomas M. Moore, Jason D. Fowlkes, and Philip D. Rack Citation: Journal of Vacuum Science & Technology A 30, 041404 (2012); doi: 10.1116/1.4731254 View online: http://dx.doi.org/10.1116/1.4731254 View Table of Contents: http://scitation.aip.org/content/avs/journal/jvsta/30/4?ver=pdfcov Published by the AVS: Science & Technology of Materials, Interfaces, and Processing Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37

Transcript of In situ laser processing in a scanning electron microscope

In situ laser processing in a scanning electron microscopeNicholas A. Roberts, Gregory A. Magel, Cheryl D. Hartfield, Thomas M. Moore, Jason D. Fowlkes, and Philip D.

Rack Citation: Journal of Vacuum Science & Technology A 30, 041404 (2012); doi: 10.1116/1.4731254 View online: http://dx.doi.org/10.1116/1.4731254 View Table of Contents: http://scitation.aip.org/content/avs/journal/jvsta/30/4?ver=pdfcov Published by the AVS: Science & Technology of Materials, Interfaces, and Processing

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37

In situ laser processing in a scanning electron microscope

Nicholas A. RobertsDepartment of Materials Science and Engineering, University of Tennessee, Knoxville, Tennessee 37996 andOmniprobe, Inc., an Oxford Instruments Company, 10410 Miller Rd., Dallas, Texas 75238

Gregory A. Magel, Cheryl D. Hartfield, and Thomas M. MooreOmniprobe, Inc., an Oxford Instruments Company, 10410 Miller Rd., Dallas, Texas 75238

Jason D. FowlkesCenter for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831

Philip D. Racka)

Department of Materials Science and Engineering, University of Tennessee, Knoxville, Tennessee 37996 andCenter for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831

(Received 10 March 2012; accepted 8 June 2012; published 29 June 2012)

Laser delivery probes using multimode fiber optic delivery and bulk focusing optics have been

constructed and used for performing materials processing experiments within scanning electron

microscope/focused ion beam instruments. Controlling the current driving a 915-nm semiconductor

diode laser module enables continuous or pulsed operation down to sub-microsecond durations,

and with spot sizes on the order of 50 lm diameter, achieving irradiances at a sample surface

exceeding 1 MW/cm2. Localized laser heating has been used to demonstrate laser chemical

vapor deposition of Pt, surface melting of silicon, enhanced purity, and resistivity via laser

annealing of Au deposits formed by electron beam induced deposition, and in situ secondary

electron imaging of laser induced dewetting of Au metal films on SiOx. VC 2012 American VacuumSociety. [http://dx.doi.org/10.1116/1.4731254]

I. INTRODUCTION

In situ laser processing in a scanning electron microscope

(SEM) and a focused ion beam (FIB) have been demon-

strated previously by Funatsu et al.1 and Hwang et al.2 where

sample heating and enhanced ion beam induced deposition,

as well as laser assisted chemical vapor deposition (LCVD),

were demonstrated, respectively. This capability, still rela-

tively limited, can offer tremendous utility for direct-write

micro- and nano-scale synthesis that is routinely performed

in a SEM and FIB system. LCVD in the SEM and FIB can

be a valuable complementary technique to focused ion and

focused electron beam induced processing. In addition, the

focused laser can be used for localized thermal treatments

(e.g., annealing), which can enable in situ observation of

phase transformations, can be coupled with energy disper-

sive x ray analysis for in situ chemical analysis for thermo-

dynamic and kinetic information, and can be used with

electron backscatter diffraction for in situ structural informa-

tion. Finally, the confluence of thermal processing and stand-

ard focused electron and ion beam induced processing could

be of significant interest as room temperature deposition typ-

ically yields high carbon containing deposits due to the mo-

lecular composition of the organometallic precursor gases

used.3–5 Some organometallic precursor gases have been

shown to form deposits that can be annealed ex situ to

achieve high purity metals.6 The goal of this work is to dem-

onstrate the capabilities and results using a laser delivery

system for localized irradiation and heating of samples in a

SEM.

In the work described here, we briefly introduce a system

used to locally deliver a sub-millimeter focused high-power

(multiple watts) laser spot to a sample inside a conventional

SEM or dual scanning electron/ion beam system. Transient

and steady state surface temperatures from room temperature

to over 2000 �C have been achieved using near-infrared light

from a semiconductor diode laser module. The delivered

laser power can be continuous, or pulsed down to sub-

microsecond durations at repetition rates of up to 100 kHz or

more by using current modulation. While previous systems

have used micromachined optical fibers that were manipu-

lated in very close proximity to the substrate,2 this lensed

probe is less invasive since the optics can be configured to

allow multiple-millimeter working distances. Thus collateral

damage to the optical system from micro/nanoscale process-

ing is minimized. Subsequently, we demonstrate some pre-

liminary functionality realized with the laser delivery probe

including laser-induced chemical vapor deposition, localized

thermal annealing of silicon (sufficient to melt the silicon),

laser post-annealing of electron-beam induced deposition of

gold nanostructures, and in situ real-time scanning electron

imaging of pulsed laser-induced dewetting of thin metal

films.

II. EXPERIMENTAL SETUP

A schematic diagram of the experimental setup used in

this work is shown in Fig. 1(a). SEM/FIBs used for the cur-

rent work included a Zeiss XB1540 CrossBeamTM

and an

FEI Nova 600 NanoLab DualBeamTM

. The laser delivery

probe was mounted on one of the ports on the SEM/FIB

chamber. Alignment of the laser delivery probe with the

electron and ion beams is achieved with six alignment

a)Author to whom correspondence should be addressed; electronic mail:

[email protected]

041404-1 J. Vac. Sci. Technol. A 30(4), Jul/Aug 2012 0734-2101/2012/30(4)/041404/6/$30.00 VC 2012 American Vacuum Society 041404-1

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37

screws located on the exterior of the port of the SEM/FIB

chamber by using the SEM image and an IR camera. The IR

camera is used for a coarse alignment with features seen in

the SEM image and the camera, and the SEM image is used

for a fine alignment by depositing a thin metal film with a

low melting point and relatively high absorption at the laser

wavelength (�10 nm Ni film in this case). In some experi-

ments, a gas injection system (GIS) from Omniprobe

(OmniGISVR

) or FEI was used, mounted on another port. The

laser delivery probe is translated along its axis under com-

puter control to focus the laser spot on the sample. The sam-

ple surface was tilted toward the laser delivery probe for

some experiments reported below, and for others was normal

to the electron beam.

Several different versions of a vacuum-sealed laser deliv-

ery probe were designed and built by Omniprobe, Inc., all

having multimode fiber delivery and miniature optics to

image the fiber end face onto the sample, with varying

(de)magnifications to control the optical spot size at the sam-

ple. This configuration combines the convenience of power

delivery using fiber optics with the versatility of bulk optics

to tailor working distance and spot size. Some versions were

capable of simultaneous microscopic sample imaging using

a digital video camera. Depending on the focal length of the

final lens, laser spots of approximately 50 and 100 lm diam-

eter on the sample were obtained at normal incidence and at

optical working distances of 3–8 mm. Additionally, the laser

spot diameter is adjustable by varying the focus position

along the probe axis. The distal end of an in vacuo laser

delivery probe containing miniature optics and having an

outside diameter near the tip of less than 3.5 mm is shown in

the photograph [Fig. 1(b)]. The main shaft of the probe is

6.35 mm in diameter.

A fiber-pigtailed 915 nm 25 W multichip diode laser mod-

ule (Oclaro, BMU25B-915-01) on a temperature-controlled

mount was driven by a 10 A quasi-cw pulsed diode laser

driver (IXYS Colorado/Directed Energy, PCX-7410) in

order to control the laser power, pulse width, and duty cycle.

The optical transmission of two versions of the laser delivery

system enabled total incident laser powers at the sample of

up to 21.4 and 18.4 W for 50 and 100 lm diameter laser

spots, respectively. As will be demonstrated, this apparatus

can achieve high irradiance at the surface of the substrate.

To avoid damage to any components within the chamber

(detectors, etc.) the laser is only operated when focused onto

the substrate and positioning of the probe was such that the

reflected (although diverging) beam is not in the path of the

detectors. The PCX-7410 driver allows for continuous wave

operation of the laser up to 12 W output and pulsed mode up

to 23 W (at up to 95% duty cycle). The programmable pulse

width range of the driver is nominally 100 ns to 475 ms;

however, the optical pulse width was measured to exactly

match the nominal setting only to just below 10 ls, and 1 ls

pulse widths resulted in ringing oscillations for a significant

portion of the pulse length (several hundred nanoseconds)

due to the internal inductance of the laser module and the

fast rise time of the current waveform. However, the rise and

fall times of the laser and drive combination were measured

to be between 20 and 30 ns, and sub-microsecond optical

pulses at full amplitude were obtainable.

III. RESULTS AND DISCUSSION

LCVD was performed using the trimethyl-(methylcyclo-

pentadienyl) platinum(IV) ((MeCp)Pt(Me3)) precursor gas,

which was locally injected using an Omniprobe OmniGIS

gas injection system. The laser assist in this case is assumed

to be pyrolytic, based on the relatively long wavelength of

the laser, which has a photon energy (<1.4 eV) insufficient

to photolytically dissociate the precursor.2,7–9 Figure 2

shows Pt deposition on a silicon substrate via an LCVD pro-

cess which was accomplished using �5.9 W (�300 kW/cm2

with a 50 lm diameter laser spot), 100 ms pulse width, and

50% duty cycle over a 100 s total exposure time. The Pt de-

posit shows a concave depression suggestive of the nonuni-

form radial temperature distribution and signs of reacting

(and possibly melting via a Pt-Si eutectic) with the underly-

ing Si substrate across the laser spot forming a platinum sili-

cide. The area of the deposit can also be varied by using

different focusing optics, or increased by defocusing the

laser spot on the sample.

Another useful capability enabled by the laser delivery

probe is in situ thermal treatment. Thermal treatments, such

as annealing and hardening, are routinely performed with

high power laser diodes.10,11 Adding this capability to a

FIG. 1. (Color online) (a) Schematic diagram of experimental setup. (b)

Close up of in vacuo laser delivery probe tip using miniature optics.

FIG. 2. Laser-assisted chemical vapor deposition of Pt on Si. EDS measure-

ments show up to 50 at. % Pt in laser-assisted chemical vapor deposition.

041404-2 Roberts et al.: In situ laser processing in a scanning electron microscope 041404-2

J. Vac. Sci. Technol. A, Vol. 30, No. 4, Jul/Aug 2012

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37

SEM or SEM/FIB system allows for imaging at the nano-

scale during the thermal treatment process. In Fig. 3 we

show examples of the heating capabilities of the laser deliv-

ery system on a Si substrate. Figure 3(a) shows SEM images

of the heat-affected zone of a Si substrate exposed to differ-

ent irradiance (top row) and pulse widths (bottom row). Fig-

ure 3(b) shows the simulated temperature evolution of the

surface of the Si substrate for different pulse widths at a

fixed laser power at the substrate surface of approximately

3.8 W (194 kW/cm2 with a 50 lm diameter laser spot). The

peak surface temperature of the substrate is a function of the

spot size, pulse width, and laser power as well as the thermal

and optical properties of the substrate (heat capacity, thermal

conductivity, and emissivity, reflectivity, absorption coeffi-

cient). Figures 3(c) and 3(d) are three-dimensional plots of

the temperature within the substrate for two different combi-

nations of pulse width and laser power. Figure 3(c) is the

three-dimensional plot at 3.8 W for a 100 ms pulse and Fig.

3(d) is a 20 W pulse for 1 ms. The peak temperatures in Figs.

3(c) and 3(d) are similar, while there is a clear difference in

the temperatures reached within the substrate which demon-

strates that the thermal penetration depth is dependent on the

pulse width. One major issue with SEM imaging during ther-

mal processing is thermal drift (motion) of the substrate,

which can cause poor image quality.1 Reducing the thermal

penetration depth and therefore the effective heating depth

and time results in less thermal drift. Ideally surface heating

of the substrate can be achieved without significant heating

of the entire substrate. Short pulse widths and high powers

produce the best results for minimized bulk substrate heating

and reduced thermal drift. As described by Kennedy et al.,10

Li,11 and Doubenskaia and Smurov,12 all of the relevant

laser processing regimes (annealing, melting, and ablation)

of pulsed and continuous wave solid state laser processing

can be achieved, as power density can easily be controlled

via a combination of laser power and focus, and the tempera-

ture can be controlled via the power density (up to

�1.2 MW/cm2 and �300 kW/cm2 for the 50- and 100-lm

diameter laser spots, respectively) and the pulse width

which, depending on the choice of laser source, can range

from tens of nanoseconds to continuous wave.

One specific area of thermal treatment of recent interest is

for the purification of EBID structures via annealing.

Annealing of EBID structures is typically performed ex situvia standard heating methods.13 Recent work by Mulders

et al. showed an improvement in EBID purity with the use

of a heated stage during the EBID process. Increasing the

substrate temperature showed an increase in purity, but with

further increases in temperature a reduced growth rate and

even thermally stimulated deposition occurred.14 An impor-

tant application for EBID in which deposit purity is of signif-

icant importance includes integrated circuit repair and

editing.4 In this case, bulk-heating methods can be detrimen-

tal to other chip components. The use of localized pulsed

laser heating could mitigate damage to thermally sensitive

areas while resulting in the same, if not better, post anneal-

ing properties. The surface temperature and heat affected

zone can be controlled by varying the laser power and pulse

width. Additionally, application specific laser spot sizes can

be used when necessary. Figure 4 shows the gold-to-carbon

ratios of these gold EBID structures on a titanium substrate

after undergoing different laser annealing treatments. The

EBID structures are 500 nm� 500 nm squares that are

approximately 250 nm thick. The structures were deposited

FIG. 3. (Color online) Demonstration of laser heating of a Si substrate. (a) Spots from laser heating of the Si substrate under different conditions—longer heat-

ing time and higher power show larger heat affected zones as well as the distinct morphology change. (b) Simulated temperature evolution of a Si substrate

surface for different pulse widths at 3.8 W of laser power. (c) Simulated temperature-depth profile of a 100 ms pulse width at 3.8 W of laser power. (d) Simu-

lated temperature-depth profile of a 1 ms pulse width at 20 W of laser power.

041404-3 Roberts et al.: In situ laser processing in a scanning electron microscope 041404-3

JVST A - Vacuum, Surfaces, and Films

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37

from dimethyl (acetylacetonate) gold(III) (Me2Au(acac))

precursor gas using an FEI GIS. The depositions were per-

formed with a 5 keV and 400 pA beam with a 10 ls dwell

time for a total of 5 min. The base chamber pressure was

below 1 � 10�6 Torr prior to depositions. In each of these

cases the laser power was adjusted from 0 (as-deposited

samples) to 234 kW/cm2, while using a constant pulse width

of �500 ns at a frequency of 100 kHz, yielding a 5%

duty cycle. The annealing process was performed for 5 min

in each case. As mentioned, these optical pulses likely

exhibited ringing associated with the system inductance as

was observed for pulse widths of less than 1 ls. The gold-to-

carbon ratios were obtained by energy dispersive x ray

spectroscopy (EDS) measurements of the deposits and the

composition estimated using the system sensitivity factors.

Also plotted in Fig. 4 are the expected gold to carbon

ratios of the precursor molecule (dimethyl-(acetylacetonate)

gold(III)) and the expected ratio when individual carbon

atoms are progressively removed from the molecule. The

gold-to-carbon ratios of the as-deposited structures agree

well with recent results by Wnuk et al.,15 who showed

through x ray photoelectron spectroscopy that one carbon

atom was removed from the molecule during the electron

beam induced dissociation. Our EDS measurements show

that an average of two additional carbon atoms are removed

from the molecule throughout the deposit with the laser

annealing process at 234 kW/cm2. The SEM images appear

to show that the surface of the structure is composed of

higher purity gold while the underlying deposit likely

remains stoichiometrically similar to the as-deposited

material.

A limitation of a laser anneal of a deposit, as with other

anneals, is that the laser can only remove carbon at or near

the surface, leaving a large concentration of carbon remain-

ing in the bulk,13 depending on the surface-to-volume ratio

of the deposit. Insets (a) and (b) of Fig. 4 show SEM images

of as-deposited and annealed gold EBID structures, respec-

tively. A morphological change, as well as a change in

contrast, is obvious between these two images. The surface

of the annealed deposit shows a collection of gold particles

where carbon atoms have been removed, leaving voids.

Based on the EDS measurements the bulk/matrix of the

deposit still contains some carbon. Longer annealing times

could potentially allow for additional carbon to diffuse out

of the deposit leaving only pure gold. Botman et al.6 also

showed a comparable reduction of carbon concentration

from approximately 85 to 25 at. % in deposits from the

dimethyl-(acetylacetonate) gold(III) precursor gas by anneal-

ing at 500 �C in oxygen. The deposits made by Botman et al.in Ref. 6 were approximately 1.8 lm� 1.8 lm and 200 nm

thick.

Gold EBID structures were also deposited onto four-point

probe structures to obtain resistivity measurements of the

deposits as an additional comparison of the purity of the de-

posit. Figure 5(a) shows the measured resistivity of laser

annealed and as-deposited gold lines across four-point probe

devices. Figures 5(b) and 5(c) show SEM images of the as-

deposited and laser annealed gold lines on the four-point

probe devices, respectively. One as-deposited structure is

compared to three differently annealed structures. The

annealed structures were exposed to the same laser condi-

tions and for the same total time. Table I describes the

annealing process for each sample measured in Fig. 5. Laser

annealing and EBID was performed for the same total time

in each of the laser annealed samples, but was performed in

a layer-by-layer fashion as noted by the number of layers. In

the layered approach, a fraction of the total gold EBID stack

was deposited and then laser annealed for the same fraction

of the total laser annealing time. The resistivity measure-

ments show a 2 orders of magnitude reduction in the resistiv-

ity of the deposit by laser annealing. These results exhibit a

greater improvement than that previously reported via stand-

ard post-annealing in various ambients.6 Additionally, the

FIG. 4. (Color online) (a) Gold to carbon ratios for annealed and as-

deposited gold EBID structures with estimated gold to carbon ratios of the

precursor molecule and incremental loss of carbon atoms. (b) SEM image of

as-deposited gold EBID structure. (c) SEM image of annealed gold EBID

structure.

FIG. 5. (Color online) (a) Resistivity and gold to carbon ratios of annealed

and as-deposited Au EBID on four-point probe structures. (b) SEM image of

the as-deposited structure on a four-point probe device. (c) SEM image of

an annealed gold EBID structure on a four-point probe device.

041404-4 Roberts et al.: In situ laser processing in a scanning electron microscope 041404-4

J. Vac. Sci. Technol. A, Vol. 30, No. 4, Jul/Aug 2012

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37

same morphological change was observed in each of the

annealed structures on the four-point probe devices as was

seen with the pads in Fig. 4.

Finally, the laser delivery system enabled us to observe

the pulsed laser induced dewetting evolution of metal films

in real-time. Based on the estimated irradiance and optical

properties, and the resultant time scale, the dewetting is

believed to occur in the solid state. Dewetting time scales in

the solid state are much slower than in the liquid state, ena-

bling the ability to observe hole formation, rim retraction,

and the formation of the resulting nanoparticles. Figure 6

shows SEM images of the dewetting process of a 10 nm Au

film on SiOx subjected to 100 ms pulse widths (20% duty

cycle) at approximately 51 kW/cm2. Figure 6(a) shows the

as-deposited Au film, Fig. 6(b) shows the film after 30 s of

pulsed laser irradiation, and Fig. 6(c) shows the film at equi-

librium after 85 s. Figures 6(d)–6(f) show higher magnifica-

tion images of the same film at longer times where a hole

has formed and the retraction of the edges is observed.16

Thin film and patterned thin film instabilities in both the

liquid17–19 and solid state20–24 are being studied as self-

directed assembly methods for organized functional nano-

particle arrays. Clearly the ability to observe the dynamics in

real time in the SEM will enhance the understanding of thin

film instabilities as well as other dewetting dynamics, which

could lead to an improved ability to manipulate and assem-

ble nanostructures. Additionally, one can envisage the bene-

fit observing in real time various thermally induced

processes exploiting other energetic electron interactions

such as EDS and wavelength dispersive x ray spectroscopy

for in situ chemical analysis, and electron backscattered dif-

fraction for in situ crystal structure analysis.

In each application of in situ laser processing the laser

power and pulse width can be varied to control the surface

temperature and penetration depth (thermal budget). Addi-

tionally, the wavelength and pulse shape of the laser can be

adjusted for the specific application and material system.

IV. CONCLUSIONS

A new laser delivery system for in situ materials process-

ing in a SEM/FIB is described, and several applications are

briefly explored and demonstrated. Specifically, laser

induced chemical vapor deposition of Pt is demonstrated

using a standard gas injection system on the dual-beam

SEM/FIB system. Laser melting of silicon as a function of

both laser power and pulse width is also shown and the peak

temperature and heating depth is illustrated as a function of

irradiance and pulse width. Laser annealed Au EBID depos-

its are shown to decrease the carbon content and to reduce

the resistivity of the deposits by approximately 2 orders of

magnitude. This is slightly improved relative to what was

previously observed after an ex situ post-deposition anneal.

Greater improvements could still be possible by performing

more anneals throughout the growth of the deposit, and

eventually implementing a synchronized in situ EBID/laser

assist process to remove carbon on a layer-by-layer level. It

is also speculated that in situ laser processing during EBIE

could also be used to produce faster etch rates and also to

eliminate redeposition of etch products.25 Finally, in situ sec-

ondary electron imaging of pulsed laser induced dewetting is

performed for a thin gold film on SiOx.

ACKNOWLEDGMENTS

The authors acknowledge that a portion of this research

was conducted at the Center for Nanophase Materials Scien-

ces, which is sponsored at Oak Ridge National Laboratory

by the Scientific User Facilities Division, Office of Basic

Energy Sciences, U.S. Department of Energy. Omniprobe

acknowledges that the laser probe development was sup-

ported in part by the National Science Foundation under

Small Business Innovation Research Grant No. IIP-0956765

to Omniprobe, and under Grant No. IIP-1059286 to the

American Society for Engineering Education.

1J. Funatsu, C. V. Thompson, J. Melngailis, and J. N. Walpole, J. Vac. Sci.

Technol. B 14, 179 (1996).2D. J. Hwang, N. Misra, C. P. Grigoropoulos, A. M. Minor, and S. S. Mao,

J. Vac. Sci. Technol. A 26, 1432 (2008).3S. J. Randolph, J. D. Fowlkes, and P. D. Rack, Crit. Rev. Solid State

Mater. Sci. 31, 55 (2006).4I. Utke, P. Hoffmann, and J. Melngailis, J. Vac. Sci. Technol. B 26, 1197

(2008).5W. F. van Dorp and C. W. Hagen, J. Appl. Phys. 104, 081301 (2008).6A. Botman, J. J. L. Mulders, R. Weemaes, and S. Mentink, Nanotechnol-

ogy 17, 3779 (2006).7W. Turney, Y. M. Hung, S. G. Starcevich, P. S. Cardinahl, V. H. Grassian,

and K. A. Singmaster, Chem. Mater. 4, 1192 (1992).8M. Wrighton, Chem. Rev. 74, 401 (1974).

TABLE I. Details of the samples measured in Fig. 5. In each case the pulse

width was 500 ns at a frequency of 100 kHz. The final deposit thickness was

250 nm in each case.

Sample

Power

(W)

Number of

anneals

Anneal time

(min)

Total anneal

time (min)

As-deposited 0 0 0 0

Laser 1 9.6 1 16 16

Laser 4 9.6 4 4 16

Laser 8 9.6 8 2 16

FIG. 6. (Color online) SEM images of pulsed laser induced dewetting of a

10 nm Au film on SiOx irradiated at 51 kW/cm2 in 100 ms pulse widths at a

20% duty cycle. (a) As-deposited film. (b) Film after 30 s of irradiation. (c)

Film after 85 s of irradiation. (d)–(f) Higher magnification images of rim de-

velopment of a hole formed in the film at longer times.

041404-5 Roberts et al.: In situ laser processing in a scanning electron microscope 041404-5

JVST A - Vacuum, Surfaces, and Films

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37

9I. P. Herman, Chem. Rev. 89, 1323 (1989).10E. Kennedy, G. Byrne, and D. N. Collins, J. Mater. Process. Technol.

155–156, 1855 (2004).11L. Li, Opt. Lasers Eng. 34, 231 (2000).12M. Doubenskaia and I. Smurov, Appl. Surf. Sci. 252, 4472 (2006).13A. Botman, J. J. L. Mulders, and C. W. Hagen, Nanotechnology 20,

372001 (2009).14J. J. L. Mulders, L. M. Belova, and A. Riazanova, Nanotechnology 22,

055302 (2011).15J. D. Wnuk, J. M. Gorham, S. G. Rosenberg, W. F. van Dorp, T. E. Madey,

C. W. Hagen, and D. H. Fairbrother, J. Appl. Phys. 107, 054301 (2010).16See Supplementary Material Document No. http://dx.doi.org/10.1116/

1.4731254 for an in situ scanning electron movie of the dewetting process.17J. Bischof, D. Scherer, S. Herminghaus, and P. Leiderer, Phys. Rev. Lett.

77, 1536 (1996).

18S. Herminghaus, K. Jacobs, K. Mecke, J. Bischof, A. Fery, M. Ibn-Elhaj,

and S. Schlagowski, Science 282, 916 (1998).19Y. Wu, J. D. Fowlkes, N. A. Roberts, J. A. Diez, L. Kondic, A. G. Gonza-

lez, and P. D. Rack, Langmuir 27, 13314 (2011).20A. L. Giermann and C. V. Thompson, J. Appl. Phys. 109, 083520 (2011).21D. Kim, A. L. Giermann, and C. V. Thompson, Appl. Phys. Lett. 95, 1903

(2009).22J. Ye and C. V. Thompson, Phys. Rev. B 82, 3408 (2010).23Y.-J. Oh, C. A. Ross, Y. S. Jung, Y. Wang, and C. V. Thompson, Small 5,

860 (2009).24R. Clearfield, J. G. Railsback, R. C. Pearce, D. K. Hensley, J. D. Fowlkes,

M. Fuentes-Cabrera, M. L. Simpson, P. D. Rack, and A. V. Melechko,

Appl. Phys. Lett. 97, 3101 (2010).25D. Burgess, P. C. Stair, and E. Weitz, J. Vac. Sci. Technol. A 4, 1362

(1985).

041404-6 Roberts et al.: In situ laser processing in a scanning electron microscope 041404-6

J. Vac. Sci. Technol. A, Vol. 30, No. 4, Jul/Aug 2012

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 129.123.121.225 On: Mon, 13 Jan 2014 14:42:37