Article

Cite This: Anal. Chem. 2019, 91, 9952−9961 pubs.acs.org/ac

Multiscale Visualization of Colloidal Particle Lens Array Mediated

Plasma Dynamics for Dielectric Nanoparticle Enhanced
Femtosecond Laser-Induced Breakdown Spectroscopy
Mengmeng Wang,† Lan Jiang,*,† Sumei Wang,†,‡ Qitong Guo,† Feng Tian,† Zhuyuan Chu,†
Jin Zhang,† Xin Li,† and Yongfeng Lu§
†
Laser Micro/Nano Fabrication Laboratory, School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081,
China
‡
Department of Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

§
Department of Electrical and Computer Engineering, University of NebraskaLincoln, Lincoln, Nebraska 68588-0511, United
Downloaded via UNIV OF ENG & TECH LAHORE on April 16, 2023 at 07:20:37 (UTC).

States

ABSTRACT: A multiscale visualization of silica colloidal particle

lens array (CPLA) assisted laser ablation of copper is investigated.
The distributed holes on a crater of CPLA-deposited Cu (CPLA-
Cu) show a near-field effect by the silica nanoparticles (NPs), and
the plasma emission signal of CPLA-Cu is 3−5 times as strong as
that of Cu. Time-resolved plasma expansion, shockwave
propagation, plasma plume emission, and nanoparticle distribution
are observed and analyzed for ablations on both Cu and CPLA-Cu
substrates. The initial expansion of plasma generated on CPLA-Cu
is faster than that of pristine Cu. The shockwave of CPLA-Cu is
rounder and its plasma plume is wider than those of Cu. The
nanoparticle distribution shows a strong lateral collision during
plume ejection for CPLA-Cu. Plasma characterization shows the
increased plasma temperature is the key reason for femtosecond
laser-induced breakdown spectroscopy (fs-LIBS) signal enhancement. This work demonstrates the signal enhancement effect of
dielectric NPs on fs-LIBS and provides insights into hydrodynamics of the fs laser-induced plasma generated on CPLA-
deposited substrate.

L aser-induced breakdown spectroscopy is of great im-

portance and popularity for elemental determination and
detection.1 Recently, the femtosecond (fs) laser has been
widely utilized as the plasma generation source attributing to
its reduced continuum radiation and minimized ablated crater
size.2,3 Due to the small plasma size, the spectroscopic signal
using a fs laser is relatively lower than that generated by a
nanosecond (ns) laser. Various methods such as focus
adjustment, ambient atmosphere conditioning, double-pulse
configuration, and pulse shaping, etc., were implemented to
improve the fs-LIBS performance, and these methods are
comprehensively reviewed by Labutin et al.3 Many of these
methods are advantageous in signal enhancement but suffer
from experimental complexity and increased instrumental
costs. Recently, a simple method reported by De Giacomo et
al. can greatly enhance the ns-LIBS signal by 1−2 orders of
magnitude through depositing noble metallic nanoparticles
(NPs) on the analytes,4 and this technique has been
investigated and applied to various samples.5−11 Thereafter,
Rusak et al. reported a SE-LIBS (surface-enhanced laser-
induced breakdown spectroscopy) method for S/N improve-
ment of the detection of aqueous solutions, namely, using
SERS (surface-enhanced Raman spectroscopy) substrates to
replace the normal slide glass.12 Dong et al. used a
superhydrophobic surface combined with magnetic confine-
ment to further improve the experimental repeatability and
detection limit.13 The mechanism of signal enhancement is
that the dropped metallic NPs augment the local electro-
magnetic field of the incident laser through their surface
plasmons and, hence, improve the efficiency of seed electron
production during laser ablation.10 On the basis of this,
metallic NPs are also employed to enhance the fs-LIBS signal.
Kalam et al. reported 2-fold enhancement on fs-LIBS of Cu
using Ag NPs,14 and again they reported a similar enhance-
ment factor on fs ST-FIBS (standoff filament induced
breakdown spectroscopy) of explosives.15 Yang et al. studied
the effect of Au NP size on the enhancement of SiO2 in a fs
double-pulse configuration and reached an enhancement factor
of 13 using Au NP only, and further to 30 with the
Received: April 6, 2019
Accepted: July 3, 2019
Published: July 3, 2019

© 2019 American Chemical Society 9952 DOI: 10.1021/acs.analchem.9b01686

Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry
Figure 1. Experimental setup for the multiscale observation (M, mirror; BS, beam splitter; PBS, polarizing beam splitter; BBO, β-BaB2O4; 1/2 λ,
half-wave plate).
combination of the second pulse reheating,16 which is the
reported highest enhancement for NPs on fs-LIBS.
Normally, NPs with a diameter less than the wavelength of
the incident beam may redistribute the electromagnetic field
through optical resonance,17 namely, plasmon resonance
(PR)18 and spherical cavity resonance (SCR).19 PR is
generated in metallic NPs due to the oscillation of the free
electrons. In the case of substrate consisting of metallic NPs,
the laser−material interaction may be altered. Inspired by the
NP-enhanced LIBS mentioned above, Chen et al. compared
the plasma properties of bulk and nano metal plates to
illustrate the influence of the NPs on plasma properties.20 On
the other hand, SCR occurs for a transparent dielectric sphere
and is sensitive to the particle size. This type of resonance is
sharp and efficient since the radiation divergence for the
resonances mode is small. These resonances can provide hot
spots in the near field around the particle for nano-
manufacturing. For example, silica micro-/nanospheres-assis-
ted laser fabrication has been widely investigated in the
formation of sub-wavelength features.21−25 It has been
reported that the use of CPLA can enhance the local light
intensity to enhance the laser ablation.26 A monolayer of
micro-/nanospheres on the sample surface acts as an array of
near-field lenses that focus the incident light into the multitude
of distinct spots.24 Previous extensive studies focused on the
nanostructures fabrication on the substrate using low laser
pulse energy; however, the strong ablation of the CPLA-
deposited substrate using high pulse energy has not been
investigated. Silica NPs are relatively low cost and may also
have a LIBS signal enhancement effect. The ultrafast laser−
matter interaction and subsequent plasma evolution is crucial
for the performance of fs-LIBS as it is the way the emission
signal occurs.27,28 Therefore, the objective of this study is to
explore the extended applicability of dielectric NPs for fs-LIBS
signal enhancement and probe the hydrodynamics of fs laser-
induced plasma generated on a CPLA-deposited substrate in a
9953
Article
temporal, spatial, and spectral manner of multiscale visual-
ization.
In this work, we used a Cu plate as the substrate and
deposited silica NPs to form a monolayer of CPLA. It is found
the deposited CPLA can enhance the emission signal of Cu by
3−5 times. To unveil the underlying mechanisms, the role of
the CPLA on the ablation process was thoroughly studied in
this work. Techniques involving reflective pump−probe
imaging, time-resolved shadowgraphy, plasma imaging, nano-
particle scattering imaging, and optical emission spectroscopy
were employed to investigate the laser-induced plasma
expansion, shockwave evolution, plume dynamics, and nano-
particle distribution during laser ablation of Cu and CPLA-Cu
substrates.
■ EXPERIMENTAL SECTION
Femtosecond Laser Direct Fabrication System. A
commercial Ti:sapphire amplifier (Spitfire Ace, Spectra
Physics) seeded by an oscillator (Maitai, Spectra Physics)
with a center wavelength of 800 nm and a pulse duration of 50
fs was used as the irradiation source. The pulse energy was
attenuated by a combination of half-wave plate and a Glan−
Thompson prism. The laser was focused by a UV fused silica
plano-convex lens (LA4380, Thorlabs) with a diameter of 25.4
mm and a focal length of 100 mm. The focused fs laser
generated air filament was monitored by a CCD equipped with
a lens tube in side view. The sample was placed on a 3D stage,
and its surface was placed 5 μm higher than the starting point
of the filament to avoid air breakdown during the laser
ablation. In this study, the pulse energy was kept to be 200 μJ
and the irradiated spot size on the sample was 45 μm.
Therefore, the laser fluence was calculated to be 12.6 J/cm2.
CCD was employed to monitor the laser processing for
keeping a fresh spot ablated by each laser pulse shot.
Deposition of Silica CPLA on Copper Surface. The Cu
surface was cleaned by ethanol solution and dried in air. A 1
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry
Figure 2. SEM images of the craters on (a) Cu and (b) CPLA-Cu; (c, d) enlarged areas on panel b.
μL drop of amorphous silica water solution (8070, Thermo
Scientific, 4.1% CV) with a particle diameter of 730 ± 20 nm
was pipetted and adhered to the Cu surface; then a 1 mL drop
of ethanol was dropped on the deposited solution to dilute the
NPs to form a silica monolayer. The NPs in the monolayer are
closely packed, and the density is calculated to be
approximately 2.1 × 108 cm−2. It should be noted that “coffee
rings” formed at the edges of the dried solution, where the NPs
aggregate to several layers or clusters. The monolayer areas in
the center were confirmed by optical microscope and scanning
electron microscope. These areas were selectively ablated
through online CCD monitoring.
Reflective Pump−Probe Imaging. As shown in Figure 1,
the time evolution of the surface structure was investigated
using reflective pump−probe imaging technique in top view.
The fs laser beam was split to two arms: one was used as the
pump to induce the ablation, and the other acting as a probe
was frequency doubled by a BBO (β-BaB2O4) crystal
(CASTECH Inc.). The pump beam irradiated on the sample
in normal direction, and the probe beam was introduced to the
incident pump beam path using a dichroic mirror. The
reflected probing light by the sample was filtered by a bandpass
filter (400 ± 10 nm). The time delay between the pump and
the probe pulse was realized using a mechanical optical delay
stage with a resolution of 16.7 fs. The ultrafast dynamic images
were taken at delay time up to 15.6 ns. Zero probe delay was
estimated within 300 fs by observing the onset of carrier
excitation and surface melting on a Si sample using a low pulse
energy of 10 μJ.29
Time-Resolved Shadowgraphy. The early stage of
plasma/shockwave expansion of laser ablation was investigated
using ultrafast pump−probe shadowgraphic imaging in side
view. Part of the probe beam was directed to propagate
through the sample surface. A CCD with a bandpass filter (400
± 10 nm) was used to image the passing profile of the probe
beam. The time delay between the pump and the probe pulse
was fixed to be the same as that in the reflective pump−probe
imaging. The zero delay was determined by the occurrence of
air breakdown.
For longer delay time, a time gated ICCD (iStar 334,
Andor) combined with a 473 nm CW laser was used to
observe the shockwave evolution. The CW laser passed
through the sample surface to the ICCD equipped with a
lens tube. A bandpass filter (470 ± 10 nm) was inserted in
front of the lens tube to block the plasma emission. The zero
delay was determined by the inception of the plasma plume
emission. Time delay was electronically adjusted by the DDG
built in the ICCD, and its time resolution was 2 ns. Each
9954
Article
shadowgraphic picture was taken for one laser bombardment
on a new surface area.
Plasma Plume Imaging. On the basis of the laser-induced
plasma emission, the time-resolved evolution of the plume was
captured by the gated ICCD. This can only show the emissive
parts in the plasma; a laser light scattering technique was
employed to observe the time evolution of the ejected matter
from the substrate. The 473 nm CW laser was tuned to be p-
polarized by a half-wave plate and irradiated on the plume, and
then light scattered by the nanoparticles, in the range of Debye
and Mie particle size,30,31 was acquired by the same
combination of lens tube equipped ICCD as that for
shadowgraphy. For each time delay, the image was acquired
for one laser shot on a fresh surface spot.
Optical Emission Spectroscopy. The plasma emits
fingerprint light when a “hot” plume expands to the air and
cools. The optical emission light was collected collinearly by
the laser focusing lens, then passed through a dichroic mirror
(10Q00UF.25, Newport), and refocused by the other identical
lens to the fiber bundle (Andor, SR-OPT-8024), which
connects to the spectrograph (Shamrock 750, Andor)
equipped with the ICCD (iStar 334, Andor). The fiber bundle
consists of 19 fiber cores, and the diameter of each is 200 μm.
The whole collection bundle is a hexagon-like arrangement and
has a maximum optics entrance of 1 mm, as shown in Figure 1.
In practice, the position of the fiber entrance was optimized by
maximizing the emission signal of the highest Cu I 515.32 line.
By moving up and down the fiber bundle position within a
range of 1 mm, the intensity of the Cu I 515.32 line changes
less than 2%. During the whole study, the position of the fiber
bundle remains unchanged. Static and time-resolved spectra
were recorded. Each spectrum was accumulated for 5 times
during acquisition, and a fresh spot was provided for each
ablation by moving the sample. The plasma temperature and
electron density were inferred by the acquired spectra.
■
RESULTS AND DISCUSSION
Crater Morphology and the Plasma Emission. The
morphology of the ablated craters as well as plasma emission
are largely influenced by the deposited CPLA. Figure 2 shows
the SEM images of the ablated craters on both Cu and CPLA-
Cu substrates. Similarly, the centers of both craters exhibit
strong redeposited droplets of the melted matters. Differently,
the outer region of the crater on Cu is composed of concentric
mound-like microstructures in Figure 2a, while small holes and
splashing matters dominate at the middle region of the crater
on CPLA-Cu shown in Figure 2c and evenly distributed small
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry Article

Figure 3. Femtosecond laser-induced breakdown spectra on Cu and CPLA-Cu substrates (the ICCD gate width is 10 μs, and the gate delay since
the arrival of the laser on the substrate is 100 ns).

Table 1. Peaks Information of the Plasma Spectra on Copper and CPLA-Cua

sp. λ (nm) gA (s−1) Eup (eV) EF (mean ± SD) sp. λ (nm) gA (s−1) Eup (eV) EF (mean ± SD)
Si I 288.16 6.51 × 108
5.08 Cu I 453.97 8.48 × 107 7.88 4.55 ± 0.044
Si I 390.55 3.99 × 107 5.08 Cu I 458.70 1.92 × 108 7.80 4.26 ± 0.016
Si II 504.10 2.80 × 108 12.52 Cu I 465.11 3.04 × 108 7.74 4.68 ± 0.077
Si II 566.96 4.00 × 108 16.39 Cu I 470.46 4.40 × 107 7.74 4.66 ± 0.021
Cu I 324.75 5.58 × 108 3.82 3.29 ± 0.097 Cu I 510.55 8.00 × 106 3.82 3.56 ± 0.042
Cu I 327.40 2.75 × 108 3.78 3.08 ± 0.035 Cu I 515.32 2.40 × 108 6.19 4.28 ± 0.045
Cu I 402.26 7.60 × 107 6.87 4.16 ± 0.081 Cu I 521.82 4.50 × 108 6.19 3.86 ± 0.043
Cu I 406.26 1.26 × 108 6.87 4.39 ± 0.016 Cu I 529.25 8.72 × 107 7.74 4.52 ± 0.033
Cu I 427.51 2.76 × 108 7.74 4.11 ± 0.031 Cu I 570.02 9.60 × 105 3.82 3.62 ± 0.011
Cu I 450.94 5.50 × 107 7.99 4.89 ± 0.055 Cu I 578.21 3.30 × 106 3.78 3.29 ± 0.040
a
“sp.”, “EF”, and “SD” correspond to “species”, “enhancement factor”, and “standard deviation” for short, respectively.

Figure 4. Time-resolved surface morphology on femtosecond laser-irradiated Cu and CPLA-Cu substrates (the frame size of each figure is 100 ×
100 μm2).

hole patterns induced by the NPs are fabricated at the outer
region as shown in Figure 2d.
Figure 3 shows the emission spectra of generated plasma on
Cu and CPLA-Cu substrates. The gate delay with respect to
the laser bombardment is 100 ns to avoid the continuum
emission (even the fs laser can minimize the continuum
radiation), and the gate width is 10 μs to collect all light
emission in whole plasma life. It can be seen from the spectra
that Si II are introduced in the CPLA-Cu plasma, indicating an
ionization of the silica NPs under the intense irradiation. The
main species in the plasma are listed, and the enhancement
factors of different Cu I peaks are compared in Table 1, in
which the mean values and the standard deviation of
9955
enhancement factors are based on 10 repeated comparison
data sets. It can be seen that the enhancement factor is 3−5
corresponding to different species, which is lower than that of
∼13 using Au NPs enhanced fs-LIBS of SiO2,16 but higher
than that of ∼2 using Ag NPs enhanced fs-LIBS of Cu14 and fs
ST-FIBS of explosives.15 The enhancement factors may be
related to the incident pulse energy, the substrate material, and
the focusing, etc. Herein, we focus on the effects of deposited
CPLA on the mechanisms of plasma hydrodynamics. From the
views of the direct visualization, time-resolved characteristics of
the plasma formation, plasma expansion, shockwave evolution,
and plume dynamics are analyzed and compared for laser-
induced ablation on Cu and CPLA-Cu substrates.
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry
Figure 5. Transient shadowgraphs of material ejection on both Cu and CPLA-Cu substrates with concentric rings(a) in the initial 1.47 ns (frame
size, 125 × 125 μm2) and (b) in the time range of several nanoseconds (frame size, 186 × 186 μm2). C and S mark the contact and shock front,
respectively.
Reflective Imaging of the Surface Structures. Figure 4
shows the time-resolved pictures of the irradiated area on Cu
and CPLA-Cu substrates, respectively. As for Cu, the central
dark area expands gradually with time in the first 1.47 ns. The
dark region indicates the starting of the ablation process.32
Since the incident beam is Gaussian distributed, the hot plasma
is first formed in the center, part of the deposited laser energy
dissipates to the edges, and then the diameter of the central
area increases. After 7.8 ns, the diameter of the dark area does
not change apparently, indicating a full laser affected area for
plasma plume ejection. Attentions should be paid to the outer
rims around the dark center, which is caused by the diffraction
of the melted hot plasma. As for CPLA-Cu, there is no central
dark region in the first 1.47 ns compared to the images on Cu.
The laser intensity is redistributed by the silica NPs due to the
Mie scattering effect.26 Part of the laser transmitted onto the
Cu surface beneath the silica NPs, while part of the laser
energy (scattered by the silica and/or focused inside the silica)
acts as the source for the silica ablation. The ablation of the
silica is not as strong as that on pure Cu shown in the upper
row due to the leaked energy to the underneath before the
delay time of 1.47 ns. At 7.8 ns, the ablation on overlaid Cu
can be inferred by the small dark holes on the reflective image.
The silica NPs act as the ablation spots on the overlaid Cu
surface; thus the material melts and ejects on each spot, and
9956
Article
small craters can be seen on the surfaces. The ablated region
appears darker with the formation of deeper and larger craters
after a delay time of 7.8 ns. The white patterns of the crater
edge on CPLA-Cu at 15.6 ns may be caused by the light
scattering of destroyed silica NPs.
Shadowgraphic Imaging of the Plasma and Shock-
wave. Figure 5 shows the shadowgraphs of plasma inception
from the substrate after the laser bombardment, and the delay
time is in accord with that in reflective pump−probe imaging
for a corresponding complementary view. In the first 0.5 ns,
growing concentric rings are observed, which have been
reported by Zhang et al., noting that these strip patterns were
caused by the diffraction of the ejected plasma plume and
indicated the cooling of the plasma.33 Thereafter, the part
adjacent to the substrate blurs and the rings disappear
gradually, indicating a transition of the plasma from hot
melted liquid to a mixture of vapor and liquid. It should be
noticed that, in this time range, the compressed air shockwave
and the boundary of generated plasma are close and clear
shockwaves cannot be observed. During a delay time to several
ns, as shown in Figure 5b, a clear separation between the
compression shockwave and the contact shockwave can be
observed. The contact front and shock front are labeled in
Figure 5b.
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry
Figure 6. Time-resolved shockwave in the time delay of tens of nanoseconds (frame size, 941 × 824 μm2).
Figure 7. (a) Time-resolved plasma expansion height in the initial 1.47 ns and (b) the subsequently generated shockwave radius generated on Cu
and CPLA-Cu substrates.
With respect to longer delay time up to tens/hundreds of ns,
the captured shadowgraph using the ICCD and a CW laser is
listed and compared in Figure 6. There is a big difference in
the shockwave shape in this time range. There is an obvious
protrusion at the globe roof of the Cu’s shockwave until 400
ns, and its shape is stretched along the vertical direction. In
contrast, the shockwave shape of CPLA-Cu is rounder even
though there is a small top tip until 80 ns and becomes a
closed circle after then. It is worth noting that the protrusion is
not produced by air breakdown. This is because air breakdown
should be monitored in advance of the shockwave if it
occurred. Moreover, since air breakdown is generated by the
focused intense laser, the position of air breakdown depends
on the laser focusing and cannot be influenced by the
shockwave. Besides, as mentioned above, the sample was
placed above the air breakdown position.
9957
Article
The plasma expansion is partially stretched in the vertical
position by the tensile stress in the material, while the
shockwave evolution is approximately a process of adiabatic
expansion. Thus, they are different in expansion mechanisms.
The height of plasma expansion and the radius of shockwave
expansion are separately described. The plasma expansion
height in the initial 1.47 ns is shown in Figure 7a; herein the
evolutions of lateral diameters are not shown since they are
very similar and mostly depend on the stretching of vertical
expansion. It can be seen that the height of the plasma front
generated on CPLA-Cu is slightly larger than that on Cu,
indicating a faster plasma plume expansion. The lateral and
vertical edges of the shock front are measured and plotted vs
time as shown in Figure 7b. The protrusion is regarded as the
boundary of the vertical edge. The data are fitted using the
Sedov−Taylor equation.34−36
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry
Figure 8. Comparison of time-resolved plasma plumes on both Cu and CPLA-Cu substrates (the dashed white line shows the substrate surface,
and the dashed white curve represents the shock front shown in Figure 6).
Figure 9. Comparison of time-resolved plume scattering on both Cu and CPLA-Cu substrates (the dashed white line shows the substrate surface,
and the dashed white curve highlights the shock front).
iEy
R = λjjjj zzzz
1/(2 + β)
k ρ{
t 2/(2 + β)
(1)
where R and E are the expansion distance and the energy
converted to the shockwave, respectively. λ, ρ, β, and t are a
constant approximately equal to unity, the mass density of the
outer atmospheric gas, propagation dimensionality, and
propagation time, respectively. It was found the expansion
along the lateral direction is the same for both Cu and CPLA-
Cu, but in the vertical direction the shockwave of Cu evolves
faster than that of CPLA-Cu with a time index of 0.45 vs 0.42.
Time-Resolved Plasma Plume Imaging. The emissive
shape of the plasma plume is shown in Figure 8. The gate
width is 10 ns, and the time step is 50 ns. It can be seen that
the plasma on CPLA-Cu is much larger than that of Cu in the
first 10 ns. After then, the plume of CPLA-Cu can be described
as two parts, the upper part is like a small water drop and cools
with time, while the lower part is wider than the upper part and
its emission lasts longer than the upper part. With respect to
Cu, the plasma plume ejects from the substrate surface
vertically and the hot front cools. By depicting the shockwave
front on the emissive plasma shape, it can be seen the front of
the ejected plume can “break” the shockwave. It can be
inferred that the protrusion of the compressed air shockwave
shown in Figure 6 is caused by the local heating of the hot
9958
Article
plume front. The vertical ejection of the plasma plume for Cu
is stronger than that of CPLA-Cu; thus the protrusion on Cu
lasts longer (>400 ns) than that of CPLA-Cu (<160 ns).
The emissive image shows the emission intensity of the
plume but not the particle distribution. Here a laser scattering
technique is used to monitor the time-resolved nanoparticle
distribution as shown in Figure 9. The gate width is 10 ns, and
the time step is 100 ns. Due to the strong continuum emission
and the reflection from the compressed shockwave, the
scattering image in the first 100 ns is not accurate and not
shown here. For Cu, the kinetic evolution is similar to the
emission image in Figure 8, showing a clear gap with low
nanoparticle density between the upper and lower parts. The
increased gap size demonstrates that the upper part travels
faster than the lower part. For CPLA-Cu, the gap is not
apparent and the nanoparticles expand all around with a
diameter about twice larger than of that of Cu plume. It was
reported that the appearance of a fast part and a slow part in
the plume was related to the thermoelastic waves during laser
ablation.33 Hundreds of GPa stress may be induced by the
laser bombardment. When it propagates to the interior of the
material, it will compress the material strong enough to induce
a repelling force between the adjacent atoms. This outward
force induces a material ejection, its inward force propagates to
the interior, and another ejection may occur after a specific
time.33 For Cu plasma, this occurs and there is a gap due to the
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry Article

Figure 10. Comparison of time-resolved (a) plasma temperature and (b) electron density of Cu and CPLA-Cu.

intermittent ejections as shown in Figure 9. However, for fwhm (full width at half-maximum) of the measured peak
CPLA-Cu, the incident fluence is mediated by the silica NPs. It (Δλmeasured) can be expressed as39

i Δλ y ij ΔλStark yz
Δλmeasured = jjjj Stark zzzz +
was investigated that the NPs can act as lenses to enhance the
jjj zzz + (Δλinst)2
local electric field intensity at the contact region to the 2

k 2 { k 2 {
substrate, but lower the field intensity at the boundary of the
(3)
silica NPs.26 In this study, demonstrated by the peak of Si II in
Figure 3, the laser fluence is strong enough to ionize the silica The slit width of the spectrograph is set to be 200 μm (equals
spheres. Multiple small ablations occur at the NP located spots. the diameter of the fiber core) for most light collection. The
The lateral collision of the ablated material (silica and Cu) is instrumental broadening (Δλinst) was measured to be 0.16 nm
stronger than that of Cu and forms a wide plume. Meanwhile, using the Hg I 546.08 nm line of a mercury lamp. And the
the ejection of plume generated on CPLA-Cu in the vertical Stark broadening (ΔλStark) is approximately proportional to the
direction is compressed. As shown in Figure 2, the central electron density40
i Ne yz
ΔλStark ≈ 2ωjjj 16 zz
concave pit on the crater of the CPLA-Cu demonstrates the

mediated laser distribution by the silica NPs.
Time-Resolved Plasma Emission. The time-resolved
spectra of element Cu in the laser-induced plasmas on Cu
and CPLA-Cu are recorded and characterized. The gate width
for each acquisition is 50 ns, and the step increase is 100 ns,
respectively. The plasma temperature is calculated using a
Boltzmann plot. Assuming the species is in LTE (local thermal
equilibrium), the plasma temperature can be calculated using
the following equation37,38
ij Iλ yz i hcN yz
lnjjj zzz = − k + lnjjjj zz
k 10 {
(4)
where ω is the electron impact parameter of the Cu I 510.55
nm line. The calculated electron density is shown in Figure
10b, and the inset shows an example of the peak fitting using a
Vogit curve. The LTE criterion is checked and fulfilled during
the calculation of the electron density. The initial increase of
electron density on CPLA may be caused by the compression
of the covering of silica plasma on the overlaid Cu plasma. The

j gA z z
E
k { k {
electron density of the plasma is very similar except that in the
k bT 4π U ( T ) (2) first 400 ns the electron density of CPLA-Cu is larger than that
of Cu.
where I, λ, Ek, A, g, N, T, and U(T) are the peak intensity, The increase of the plasma temperature is recognized as the
wavelength, the upper energy level of the transition, the main reason for the signal enhancement. Based on the
transition probability, the statistical weight of the upper level, Boltzmann equation, the intensity enhancement can be
the total number density, plasma temperature, and the expressed by the following equation7
partition function, respectively. kb, h, and c correspond to
Boltzmann constant, Planck constant, and the speed of light in ICPLA‐Cu N U (TCPLA‐Cu)
= CPLA‐Cu

ji E ij 1 1 yzzzyz
vacuum, respectively. By plotting ln(Iλ/gA) vs Ek using ICu NCu U (TCu)

× expjjjj− k jjj zzz

different peaks, the plasma temperature can be calculated
using the slope of the fitted line. In this study, the used peaks
are 510.55, 515.32, and 521.82 nm. The calculated plasma
temperature is shown in Figure 10a, and the inset shows an
example of the Boltzmann plot for the temperature calculation
of Cu plasma at a 500 ns delay. The temperature of CPLA-Cu
is higher than that of Cu in the whole calculated time range.
The electron density is inferred through Starks broadening
of the Cu I 510.55 nm peak. The measured peak was fitted
using a Vogit profile, which is a convolution of a Gaussian
profile mainly induced by instrumental broadening and a
Lorentzian profile mainly induced by Stark broadening. The
9959
TCu zz{z{
j
k k b k TCPLA‐Cu
−
(5)
where the subscripts indicate the generated plasmas on Cu and
CPLA-Cu substrates, respectively. It shows the enhancement is
proportional to the exponent of the upper energy for a given
temperature increase. As shown in Figure 11, a linear fitting
with an adjusted R2 of 0.84 is plotted for describing the
correlation between the upper level energy and the logarithm
of the enhancement factor. The 95% confidence band (pink
area) and the prediction band (blue area) are labeled in the
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry
Figure 11. Enhancement factor of the emission lines with respect to
the upper level energy.
figure. The linear trend demonstrates the temperature increase
of the plasma is the key factor for the signal enhancement.
■ CONCLUSION
The fs laser-induced ablation on Cu and CPLA-Cu is
investigated and compared in this work. SEM images of the
ablated craters illustrate the silica NPs can enlarge the ablation
diameter with multiple holes generation at the outer region
due to the near-field effect of silica NPs. The plasma emission
shows a 3−5 times signal enhancement by the introduction of
CPLA, and the appearance of Si II lines shows an ionization of
silica during ablation. The ablation process is directly visualized
using various techniques in multiscale. Reflective pump−probe
imaging is employed to observe the modification of formed
surface structures in the first several ns. This shows that the
diameter of formed plasma increases with time and starts
ejecting in less than 10 ns for Cu, while the plasma of melted
silica NPs on the CPLA-Cu is first observed and then the small
holes induced by the lens effect of NPs on overlaid Cu are
shown on the surface. Time-resolved expansion of plasma and
shockwave are visualized by employing shadowgraphic imaging
techniques. It is observed that the plasma generated on CPLA-
Cu substrate expands faster in the vertical direction than that
on Cu in the first 1.47 ns. The stretched shockwave on Cu
substrate has an obvious protrusion lasting for more than 400
ns, while the round shockwave on CPLA-Cu has a small tip
lasting less than 160 ns. Time-resolved emissive plasma
ejection was monitored using an ICCD, showing that the
generated plume on CPLA-Cu is wider but shorter than that
on the Cu substrate. This difference indicates the vertical
plasma ejection on CPLA-Cu is compressed by the
introduction of NPs. By comparing the plasma image and
the shock front, it is inferred that the protrusion/tip is induced
by local plume front heating of the compressed air. The
distribution of ejected nanoparticles is imaged using a
scattering technique, showing a clear gap between the fast
and slow parts of the plasma generated on Cu, but this
phenomenon is not apparent on the CPLA-Cu substrate. The
time-resolved emissive plasma is characterized by calculating
its temperature and electron density. The electron density is
similar for both substrates but the temperature for CPLA-Cu is
higher than that on Cu, demonstrating the temperature
increase is the main reason for LIBS signal enhancement.
9960
Article
These visualizations provide direct insight into the fs laser
ablation process and prove a new method for fs-LIBS
enhancement using dielectric silica NPs.
■ AUTHOR INFORMATION
Corresponding Author
*E-mail: jianglan@bit.edu.cn.
ORCID
Mengmeng Wang: 0000-0003-4393-1393
Lan Jiang: 0000-0003-0488-1987
Xin Li: 0000-0002-4743-5509
Yongfeng Lu: 0000-0002-5942-1999
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
We gratefully acknowledge the financial support of the
National Key R&D Program of China (Grant No.
2017YFB1104300), the National Natural Science Foundation
of China (NSFC; Grant No. 51575053), and The Initiative
Postdocs Supporting Program (Grant No. BX20180041).
■
REFERENCES
(1) Noll, R., Laser-induced breakdown spectroscopy. In Laser-
Induced Breakdown Spectroscopy: Fundamentals and Applications;
Springer: 2012; pp 7−15, DOI: 10.1007/978-3-642-20668-9_2.
(2) Gurevich, E.; Hergenröder, R. Appl. Spectrosc. 2007, 61 (10),
233A−242A.
(3) Labutin, T. A.; Lednev, V. N.; Ilyin, A. A.; Popov, A. M. J. Anal.
At. Spectrom. 2016, 31 (1), 90−118.
(4) De Giacomo, A.; Gaudiuso, R.; Koral, C.; Dell’Aglio, M.; De
Pascale, O. Anal. Chem. 2013, 85 (21), 10180−10187.
(5) Ohta, T.; Ito, M.; Kotani, T.; Hattori, T. Appl. Spectrosc. 2009,
63 (5), 555−558.
(6) De Giacomo, A.; Dell’Aglio, M.; Gaudiuso, R.; Koral, C.;
Valenza, G. J. Anal. At. Spectrom. 2016, 31 (8), 1566−1573.
(7) De Giacomo, A.; Gaudiuso, R.; Koral, C.; Dell’Aglio, M.; De
Pascale, O. Spectrochim. Acta, Part B 2014, 98, 19−27.
(8) De Giacomo, A.; Koral, C.; Valenza, G.; Gaudiuso, R.;
Dell’Aglio, M. Anal. Chem. 2016, 88 (10), 5251−5257.
(9) Dell’Aglio, M.; Alrifai, R.; De Giacomo, A. Spectrochim. Acta, Part
B 2018, 148, 105−112.
(10) Dell’Aglio, M.; Alrifai, R.; De Giacomo, A. Spectrochim. Acta,
Part B 2018, 148, 105−112.
(11) Wen, X.; Lin, Q.; Niu, G.; Shi, Q.; Duan, Y. Appl. Opt. 2016, 55
(24), 6706−6712.
(12) Rusak, D.; Anthony, T.; Bell, Z. Rev. Sci. Instrum. 2015, 86
(11), 116106.
(13) Dong, D.; Jiao, L.; Du, X.; Zhao, C. Chem. Commun. 2017, 53
(33), 4546−4549.
(14) Kalam, S. A.; Murthy, N. L.; Krishna, J. R.; Srikanth, V. V. S. S.;
Rao, S. V. Nanoparticle Enhanced Laser Induced Breakdown Spectros-
copy with Femtosecond Pulses, Paper Th3A.89, 13th International
Conference on Fiber Optics and Photonics, Kanpur, India, Dec. 4−8,
2016; Optical Society of America: Washington, DC, USA, 2016;
DOI: 10.1364/PHOTONICS.2016.Th3A.89.
(15) Kalam, S. A.; Rao, S. V. OSA Continuum 2019, 2 (3), 554−562.
(16) Yang, F.; Jiang, L.; Wang, S.; Cao, Z.; Liu, L.; Wang, M.; Lu, Y.
Opt. Laser Technol. 2017, 93, 194−200.
(17) Jia-Jie, D.; Guang-De, C.; Cong, X.; Fan, Z. Y.; Jin-She, Y. Chin.
Phys. 2003, 12 (1), 100−106.
(18) Jain, P. K.; Huang, X.; El-Sayed, I. H.; El-Sayed, M. A.
Plasmonics 2007, 2 (3), 107−118.
(19) Decker, M.; Staude, I. J. Opt. 2016, 18 (10), 103001.
(20) Chen, A.; Jiang, Y.; Wang, T.; Shao, J.; Jin, M. Phys. Plasmas
2015, 22 (3), 033301.
DOI: 10.1021/acs.analchem.9b01686
Anal. Chem. 2019, 91, 9952−9961
Analytical Chemistry Article

(21) Luk’Yanchuk, B. S.; Wang, Z. B.; Song, W. D.; Hong, M. H.

Appl. Phys. A: Mater. Sci. Process. 2004, 79 (4), 747−751.
(22) Wu, H.; Odom, T. W.; Whitesides, G. M. Adv. Mater. 2002, 14
(17), 1213−1216.
(23) Afanasiev, A.; Bredikhin, V.; Pikulin, A.; Ilyakov, I.; Shishkin,
B.; Akhmedzhanov, R.; Bityurin, N. Appl. Phys. Lett. 2015, 106 (18),
183102.
(24) Bityurin, N.; Afanasiev, A.; Bredikhin, V.; Alexandrov, A.;
Agareva, N.; Pikulin, A.; Ilyakov, I.; Shishkin, B.; Akhmedzhanov, R.
Opt. Express 2013, 21 (18), 21485−21490.
(25) Zhang, Z.; Geng, C.; Hao, Z.; Wei, T.; Yan, Q. Adv. Colloid
Interface Sci. 2016, 228, 105−122.
(26) Pikulin, A.; Afanasiev, A.; Agareva, N.; Alexandrov, A. P.;
Bredikhin, V.; Bityurin, N. Opt. Express 2012, 20 (8), 9052−9057.
(27) Harilal, S. S.; Brumfield, B. E.; Cannon, B. D.; Phillips, M. C.
Anal. Chem. 2016, 88 (4), 2296−2302.
(28) Hou, H.; Mao, X.; Zorba, V.; Russo, R. E. Anal. Chem. 2017, 89
(14), 7750−7757.
(29) Downer, M. C.; Fork, R. L.; Shank, C. V. J. Opt. Soc. Am. B
1985, 2 (4), 595−599.
(30) Koch, J.; Schlamp, S.; Rösgen, T.; Fliegel, D.; Günther, D.
Spectrochim. Acta, Part B 2007, 62 (1), 20−29.
(31) Koch, J.; Heiroth, S.; Lippert, T.; Günther, D. Spectrochim. Acta,
Part B 2010, 65 (11), 943−949.
(32) Sokolowski-Tinten, K.; Bialkowski, J.; Cavalleri, A.; Boing, M.;
Schueler, H.; von der Linde, D. Dynamics of femtosecond-laser-
induced ablation from solid surfaces. Proceedings, High-Power Laser
Ablation, Sante Fe, NM, USA; SPIE, 1998; Vol. 3343, DOI: 10.1117/
12.321593.
(33) Zhang, N.; Zhu, X.; Yang, J.; Wang, X.; Wang, M. Phys. Rev.
Lett. 2007, 99 (16), 167602.
(34) Sedov, L. I. Similarity and dimensional methods in mechanics,
10th ed.; CRC Press, 1993.
(35) Porneala, C.; Willis, D. A. J. Phys. D: Appl. Phys. 2009, 42 (15),
155503.
(36) Zeng, X.; Mao, X. L.; Greif, R.; Russo, R. E. Appl. Phys. A:
Mater. Sci. Process. 2005, 80 (2), 237−241.
(37) Aragón, C.; Aguilera, J. A. Spectrochim. Acta, Part B 2008, 63
(9), 893−916.
(38) Zheng, L.; Kulkarni, P.; Birch, M. E.; Deye, G.; Dionysiou, D.
D. Aerosol Sci. Technol. 2016, 50 (11), 1155−1166.
(39) Torres, J.; Jonkers, J.; Sande, M. J. v. d.; Mullen, J. J. A. M. v. d.;
Gamero, A.; Sola, A. J. Phys. D: Appl. Phys. 2003, 36 (13), L55−L59.
(40) Zmerli, B.; Ben Nessib, N.; Dimitrijević, M. S.; Sahal-Bréchot,
S. Phys. Scr. 2010, 82 (5), 055301.

9961 DOI: 10.1021/acs.analchem.9b01686

Anal. Chem. 2019, 91, 9952−9961