View Article Online

View Journal

Nanoscale
Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: E. Mitsai, M.
Aouassa, L. Hassayoun, D. Stozhenko, A. Mironenko, S. Bratskaya, S. Juodkazis, S. Makarov and A.
Kuchmizhak, Nanoscale, 2019, DOI: 10.1039/C9NR01837A.

Volume 8 Number 1 7 January 2016 Pages 1–660 This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Nanoscale Accepted Manuscripts are published online shortly after
www.rsc.org/nanoscale

acceptance, before technical editing, formatting and proof reading.

Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.

You can find more information about Accepted Manuscripts in the

author guidelines.

Please note that technical editing may introduce minor changes

to the text and/or graphics, which may alter content. The journal’s
ISSN 2040-3364 standard Terms & Conditions and the ethical guidelines, outlined
PAPER
in our author and reviewer resource centre, still apply. In no
Qian Wang et al.
TiC2: a new two-dimensional sheet beyond MXenes
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.

rsc.li/nanoscale
 Page 1 of 8 Nanoscale
View Article Online
DOI: 10.1039/C9NR01837A

Jour
nalName

Si1−x Gex nanoantennas with tailored Raman response

and light-to-heat conversion for advanced sensing
applications¶
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

E. Mitsai,a M. Aouassa,b L. Hassayoun,b D. Storozhenko,a A. Mironenko,c

S. Bratskaya,c,d S. Juodkazis,e, f S. Makarov,g A. Kuchmizhaka,d†

Nanoscale Accepted Manuscript

Active light-emitting all-dielectric nanoantennas recently have demonstrated a great potential to be
highly efficient nanoscale light sources owing to their strong luminescent and Raman responses.
However, their large-scale fabrication faces a number of problems related to productivity limits of
existing lithography techniques. Thus, high-throughput fabrication strategies allowing in a facile
way to tailor of the nanoantennas emission and thermal properties in the process of their fabri-
cation are highly desirable for various application. Here, we propose a cost-effective approach
to large-scale fabrication of Si1−x Gex alloyed Mie nanoresonators possessing enhanced inherent
Raman response which can be simply tailored via tuning Ge concentration. Moreover, by tailoring
relative Ge composition one can gradually change a complex refractive index of the produced
Si1−x Gex alloy, which affects the ratio between radiative and nonradiative losses in Si1−x Gex
nanoantennas, being crucial for optimization of their optical heating efficiency. Composition-
tunable Si1−x Gex nanoantennas with optimized size, light-to-heat conversion and Raman re-
sponse are implemented for non-invasive sensing of 4-aminothiophenol molecules with a tem-
perature feedback modality and high subwavelength spatial resolution. The results are important
for an advanced multichannel optical sensing, providing an information on analyte’s composition,
analyte-nanoantenna temperature response and spatial position.

1 Introduction
species (emitters) at low concentrations using surface-enhanced
Recently, non-plasmonic nanostructures made of high refractive spectroscopy techniques. Even though Raman scattering from the
index (high-n) dielectrics (as Si, Ge, TiO2 and diamond) have emitters directly attached (without using the dielectric spacer) to
appeared as a promising all-in-one biosensing platform 1,2 . Be- plasmonic nanoantennas can be significantly enhanced, the flu-
ing designed in a proper way, such optically resonant nanostruc- orescence signal is simultaneously quenched via coupling of the
tures as well as their arrangements exhibit excitation of various emitter to non-radiative modes of the plasmonic structures 4 . Uti-
electric- and magnetic-type Mie resonances resulting in signifi- lization of bare all-dielectric antennas provides the way to real-
cant enhancement and localization of the electromagnetic fields 3 , ize the dual-sensing scheme simultaneously enhancing the fluo-
which are crucial for optical detection of various molecular rescence and Raman scattering of the attached emitters without
the need to use nanometer-thick dielectric spacers 5–13 . More-
over, comparing to plasmonic nanostructures their all-dielectric
a
Institute of Automation and Control Processes, Far Eastern Branch, Russian Academy counterparts also offer low chemical invasiveness of the analyte
of Sciences, Vladivostok 690041, Russia
b
molecules canceling any plasmon-driven chemical reactions and
Laboratory of Micro-Opto-electronic and Nanostructures (LMON), Department of
Physics, Faculty of Sciences, 5019 Monastir, Tunisia
processes which could result in misinterpretation of the mea-
c
Institute of Chemistry, Far Eastern Branch, Russian Academy of Sciences, Vladivostok surement results 14 . Other advantages of such non-plasmonic
690022, Russia high-n crystalline nanostructures include their inherent Raman
d
Far Eastern Federal University, Vladivostok 690090, Russia bands, with their yield resonantly enhanced upon excitation of
e
Swinburne University of Technology, John st., Hawthorn 3122, Victoria, Australia
f
the magnetic-type Mie modes 15 . This enhanced signal also can
Melbourne Centre for Nanofabrication, ANFF, 151 Wellington Road, Clayton, VIC
3168, Australia be used for precise in situ mapping of the local temperature of
g
ITMO University, 197101 St. Petersburg, Russia the nanostructure-analyte system in the process of SERS measure-
¶ Electronic Supplementary Information (ESI) available: [please find attached Sup- ments via tracing the thermal-induced spectral shift of the Raman
plementary information file]. See DOI: 10.1039/b000000x/
bands associated with the nanostructure’s lattice vibration 16–19 .
† E-mail: alex.iacp.dvo@mail.ru

J
our
nal
Name,
[yea
r][
,vol
.
], 1–7 | 1
 Nanoscale Page 2 of 8
View Article Online
DOI: 10.1039/C9NR01837A

This provides a facile way to maintain thermal stability of the Incident

SERS measurements allowing to avoid the overheating (or irre- radiation
versible modification) of both the nanostructures and the analyte
Temperature-
feedback SERS
molecules 18,20 .
signal from from 4-ATP
All-dielectric nanostructures supporting strong near-field en-
SiGe analyte
hancement in their hot-spots are typically fabricated in the form
of dimers using expensive and time-consuming lithography meth-
resonator
ods 6,21–23 . This issue substantially hampers widespread imple-
mentation of such structures in routine SERS experiments, which
require utilization of either a recyclable or a fresh substrate for
each subsequent measurement 24–27 . In this respect, the search
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

for easy-to-implement, reproducible and high-productive tech- Si1-xGex

4-ATP
niques allowing to fabricate high-n optically resonant nanostruc- 4-ATP
tures with outstanding characteristics over mm-scale area is of
a

Nanoscale Accepted Manuscript

great importance 28 .

Number of nanoparticles
In this paper, we present a versatile biosensing platform based e f

Base ellipticity
Side area
on the densely-arranged fully alloyed Si1−x Gex Mie nanores- 1.4
1.2
onators produced over the mm-scale substrate via solid-state 1.0
1 μm
dewetting of commercially available SOI substrate in high vac-

Middle area
uum accompanied by deposition of Ge monolayers (MLs). The d
100 200 300 400
produced hemi-spherical Si1−x Gex nanoparticles demonstrate Si1-xGexnanoparticle diameter D, nm
pronounced optical response which can be tuned by controlling
the relative concentration of Ge. Specifically, incorporation of
Central area
1 μm
g 30 nm

Ge allows to tailor the complex refractive index of the produced c Si1-xGex

Si1−x Gex nanoparticles, which affects the ratio between radiative 0.2 0.1 0
x, rel.units D
and nonradiative losses in such Mie nanoresonators, being crucial
for optimization of their optical heating efficiency. Composition-
b 1 μm
SiO2
Si substrate
tunable Si1−x Gex nanoparticles with optimized size, calibrated
Fig. 1 Si1−x Gex Mie nanoresonators produced via spontaneous
light-to-heat conversion and SERS enhancement are implemented
dewetting of SOI substrate in high vacuum accompanied by Ge
for non-invasive sensing of 4-aminothiophenol molecules with a deposition. (a) Artistic representation of the alloyed Si1−x Gex
temperature feedback modality and high subwavelength spatial nanoparticles with 4-ATP molecules adsorbed on their surface. Isolated
resolution. The obtained results can find an application in de- Mie nanoresonator is pumped with laser radiation resulting in SERS
signing next-generation multichannel sensing platforms for truly from the attached molecules as well as Stokes Raman yield related to
Si1−x Gex lattice phonons, which provides information on the local
non-invasive and highly stable optical detection of biomolecules,
temperature of the molecule-nanoantenna system and corresponding
pollutants, explosives, chemical reaction, etc., providing informa- spatial position the nanoresonator. (b) Representative top-view SEM
tion on analyte‘s composition, analyte-nanoantenna temperature image showing three characteristic areas on the dewetted SOI
response and their spatial position within a single measurement substrate. Colored curves indicate approximate borders between these
areas. (c-e) Close-up top-view SEM images of each area. (f)
cycle.
Distribution of base diameters D of Si1−x Gex nanoresonators in the
middle (green) and side (yellow) areas of the sample. The average base
2 Results and Discussions
ellipticity is shown by the green markers. (g) TEM image of the
In our work, we used the mm-scale biosensing substrates con- representative 220-nm diameter Si1−x Gex nanoparticle from the middle
taining densely arranged Si1−x Gex Mie resonators, which are area. The ideal circle contour is shown by the dashed purple curve.
schematically illustrated in Fig. 1a. Such nanostructures are pro-
duced via spontaneous solid-state dewetting of a SOI substrate in
the ultra-high vacuum beam epitaxy reactor (UHV-MBE) accom- low magnification. Noteworthy, the deposition of Ge initiated af-
panied by deposition of the Ge monolayers 29 to boost and modify ter the central areas with faceted crystalline Si nanoparticles ap-
the dewetting dynamics (see Methods section, for details). Note- peared (Fig. 1(b,c)) strongly perturbs the dewetting dynamics, as
worthy, after heating at 800o C for 2 hours, subsequent Ge depo- it was previously reported in 29–31 . In general, the deposition of
sition procedure and cooling one could commonly identify three Ge provide the formation of fully alloyed Si1−x Gex nanoparticles
characteristic areas on the SOI surface: (i) central area, where in the middle and side areas, which preserve rather dense ar-
the dewetting process initiated by impurities and defects in SiO2 rangement and narrow size distribution (see Fig. 1(d,e,f)). High-
layer proceeds along the crystallographic orientations (Fig. 1(b,c) resolution EDX mapping of the Si and Ge content and distribution
and Fig. S1 in Supporting information), as well as a (ii) middle within the isolated nanoparticles was previously performed in de-
and (iii) side areas containing randomly arranged nanoparticles tails in Ref. 29 confirming the homogeneous distribution of the
(Fig. 1d and 1e, respectively). All these areas on the sample sur- elements within.
face can be easily identified even with an optical microscope at Specifically, in the middle area, the density of Si1−x Gex nanos-

2| J
our
nal
Name,
[yea
r][
,vol
.
],
1–7
 Page 3 of 8 Nanoscale
View Article Online
DOI: 10.1039/C9NR01837A

tructures reaches up to 6 nanoparticles per µm2 , which main- <Ein2>

a ε
II
2
tain smooth hemi-spherical shape with an average base diame- E0

Ge concentration x, rel.units
0.24 D=190 nm
ter of D=220±30 nm and a vertical (height-to-diameter) aspect 0.8
0.20
ratio about 0.6±0.1 (Fig. 1f). TEM image of representative iso- D=220 nm
lated Si0.8 Ge0.2 nanoparticle from the middle area is shown in 0.16 0.6
Fig. 1f confirming the discussed geometrical parameters. Note- D=280 nm
0.12
worthy, the relative Ge composition x of the produced nanostruc-
0.4
tures varies with the distance from the central area (see Fig. 1b) 0.08

reaching its maximal value of 0.22 near the boundary between

0.04
0.2
central and middle areas and gradually decreasing towards side Inside
area down to 0.07. The corresponding chemical composition of 0.00
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

180 200 220 240 260 280

the alloyed nanoparticles is confirmed with Raman characteriza- Nanoparticle diameter, nm
tion (see details in Fig. S1 in Supporting information).
Optical properties of Si1−x Gex nanoparticles. At such char-
b 0.36
Outside 4 4

Ge concentration x, rel.units
Eout /E0

Nanoscale Accepted Manuscript

0.32
acteristic dimensions, the produced isolated Si1−x Gex nanoparti-
0.28 4000
cles can support electric- and magnetic-type Mie resonances in
the visible spectral range as it was previously highlighted in 31,32 . 0.24 3500

Under resonant excitation, both the EM fields near and inside 0.20 3000

such nanoresonator can be enhanced significantly. While the 0.16 2500

outer EM field will expectedly boost the fluorescence/Raman sig- 0.12 2000

nal of the molecules and emitters attached to the nanoparticle 0.08 1500

surface, the field localized inside the resonator will increase the 0.04 1000

intensity of the Raman scattering related to the lattice phonons 500

0.00
180 200 220 240 260 280
of the nanoparticle material as well as light-to-heat conversion
efficiency (see Fig. 1a). In other words, under resonant excita- Nanoparticle diameter, nm
tion the incident laser radiation can be efficiently absorbed by the
Fig. 2 Calculated EM response of Si1−x Gex nanoresonators. (a)
nanoresonator resulting in their strong heating. In this respect, 2 >/E 2 )·ε 00 (x) spatially averaged
Squared electric field amplitude (<Ein 0
thermal stability of SERS measurements with the all-dielectric over the corresponding volume of the Si1−x Gex alloyed nanoparticle and
nanostructures can not be taken for granted and should be care- multiplied over the Ge-composition dependent imaginary part of
fully controlled as it was recently discussed in 33,34 . permittivity ε 00 (x) versus the base diameter D and Ge concentration x at
532-nm pump. (b) Maximal value of Eout 4 /E 4 calculated near the surface
In the visible spectral range, Ge can be characterized by an 0
of isolated Si1−x Gex nanoparticle as a function of its base diameter D
order of magnitude larger imaginary part of permittivity being
and relative Ge concentration x at 532-nm pump. Insets in (a) show the
compared to Si 35 (see Fig. S2 in Supporting Information). In this 2D E-field distributions (E/E0 ) inside Si0.93 Ge0.07 nanoparticles of
respect, by tuning the Ge composition x in the alloyed Si1−x Gex various base diameters D. The color bar limits differ for each image to
nanoparticles one could efficiently tailor their absorption (light- show the characteristic field structure inside each nanoparticle.
to-heat conversion) efficiency. Moreover, the melting temperature
Tmelt of the Si1−x Gex alloy will also decrease linearly with Ge con-
tent increase according to 36 . For our nanostructures, the melting value. At high Ge content 37 , the heating efficiency drops as the
temperature reduces to Tmelt ≈1500 K at the highest Ge concen- 2 >/E 2
characteristic EM field intensity inside the nanoparticle <Ein 0
tration x ≈0.22 (Tmelt ≈1687 K for pure Si). tends to zero. Noteworthy, at low Ge content the maximal ef-
To shed light onto the structure of the EM fields near and in- ficiency is achieved for nanoparticles with their base diameter
side the alloyed Si1−x Gex nanoresonators of various base diam- D≈220±15 nm, which can be explained in terms of excitation of
eter D and chemical composition x as well as to access the re- the magnetic-type quadrupolar (MQ) resonance (at D≈220 nm)
lated light-to-heat conversion effects, we have performed numer- and particularly to the first-order electric-type dipolar (ED) res-
ical analysis using commercial software (Comsol Multiphysics, onance (at D≈190 nm), which is confirmed by calculating char-
see Methods). Figure 2a gives the squared electric field ampli- acteristic EM field distribution inside the Mie nanoresonator of
tude (<Ein 2 >/E 2 )·ε 00 (x) spatially averaged over the correspond- variable size (see inset in Fig. 2b). Increase of the relative Ge
0
ing volume of the Si1−x Gex resonator and multiplied over the Ge- concentration x introduces Ohmic losses which appear to affect
composition dependent imaginary part of permittivity ε 00 (x) versus strongly higher order optical resonances. This explains the grad-
the base diameter D and Ge concentration x at 532-nm excita- ual shift of the heating efficiency towards smaller nanoparticle D
tion. This calculated value is proportional to the absorbed laser at increased x, resulting in maximal light-to-heat conversion at
power and provides direct estimation of light-to-heat conversion ED resonance excitation for Ge-rich resonators.
efficiency in the alloyed nanoparticles. The presented data indi- 4 /E 4 (Fig. 2b) representing the forth
Further, we calculated Eout 0
cates that in the range of relatively Ge concentration realized in power of the maximal electric field amplitude near the surface
our sample (x<0.22), the nanoparticle heating efficiency expect- outside Si1−x Gex nanoresonator. This value allows to roughly ac-
edly increases with x owing to the corresponding increase of ε 00 (x) cess the typical enhancement factor (EF) of the SERS signal com-

J
our
nal
Name,
[yea
r][
,vol
.
], 1–7 | 3
 Nanoscale Page 4 of 8
View Article Online
DOI: 10.1039/C9NR01837A

Raman shift, cm-1 30 1000

250 350 450 550 25 c D=280 nm
.2 750
a

-1
Raman intensity, rel.units
x=0
D=250 nm

substrate
20

Δ, cm
Nanoparticle

ΔT, К
Si-Si D=210 nm

Si
15 500
Ge-Ge Si-Ge 7
10 x=0.0
250
r
300 K 5
I, mW/μm2 se g D=210 nm D=250 nm D=280 nm
la atin 0 0
0.15 e
h 0 3 6 9 12 15 18
1.5
4.0 870 K Pump intensity I, mW/μm2
7.5 Optical image SEM image Raman image

(502-507 cm-1)
14.0
d e f

Si-Si band
b
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

460 470 480 490 500 510 520 530

Raman shift, cm-1 1 μm

Nanoscale Accepted Manuscript

Fig. 3 Composition tunable light-to-heat conversion in Si1−x Gex Mie nanoresonators. (a) Raman spectra of the isolated 230-nm diameter
Si0.8 Ge0.2 nanoparticle. The main Raman bands associated with nanoparticle lattice phonons are highlighted by color. (b) Series of Raman spectra
shown in the vicinity of Si-Si Raman band obtained from similar nanoparticle at increased pump laser intensity I. (c) Spectral shift of the Si-Si Raman
band ∆ and the average nanoparticle temperature increase ∆T versus the pump laser intensity I measured for Si1−x Gex of variable diameter D and
relative Ge concentration of x=0.07 and 0.2. Corresponding temperature increase was recalculated according to the data from 16,17 . (d,e) Optical and
reference SEM images of the surface area containing densely arranged Si0.8 Ge0.2 nanoparticles. (f) Raman intensity map of the same surface area
measured in spectral range of 502–507 cm−1 .

ing from the emitters attached to the Si1−x Gex nanoparticle sur- The typical Stokes room-temperature Raman spectrum measured
face. As seen, the maximal EM field enhancement is achieved for from the isolated Si0.8 Ge0.2 resonator is shown in Fig. 3a reveal-
Si1−x Gex resonators with the diameters D≈225±15 nm (around ing four main optical bands. The first band at 520.7 cm−1 is
50% of all nanostructures in the middle and side area). This fea- related to first-order optical phonons of the Si substrate, while
ture can be fully explained in terms of excitation of the MQ res- the other lower frequency modes – to Si-Si (≈505 cm−1 ), Si-
onance in the Si1−x Gex nanoparticles, as it was mentioned above Ge (≈400 cm−1 ) and Ge-Ge (≈300 cm−1 ) nanostructure lattice
(see inset in Fig. 2b). Also, increase of the Ge content x expect- vibrations. Si-Ge and Ge-Ge Raman bands (highlighted by red
edly reduces the maximal field enhancement (Q-factor of the Mie color in Fig. 3a) have rather low intensity and large shape asym-
resonator) via increased Ohmic losses as Q ∼ 1/Im(ε). Specif- metry complicating their detection and tracing, especially at low
4 /E 4 value drops from 3700 to 1800 at increasing
ically, the Eout relative content of Ge x. The most-intense Si-Si band of alloyed
0
relative Ge concentration x from 0.07 to 0.22. In particular, this nanoparticle (highlighted by blue color in Fig. 3a) in the case of
indicates that nonradiative losses in the nanoresonator reduce the strain-relaxed material (see Fig. S4 in the Supporting Informa-
amount of accumulated light energy in its near-field, while more tion) has a spectral position which linearly scales with Ge com-
energy is absorbed in its volume. position x according to simple linear relation 17,38 ωSi−Si =520.7–
68·x (see also Fig. S1 in Supporting Information). In this respect,
Noteworthy, systematic analysis of the data presented in Fig. 2
detecting the spectral position of Si-Si Raman band provides in
helps to identify the optimal set of parameters related to ge-
situ identification of the actual Ge content in the probed Si1−x Gex
ometry and composition of the Mie resonators, which allow to
nanoresonator allowing also to distinguish the contribution of the
manage light-to-heat conversion efficiency versus the outer EM
nanostructure and underlying c-Si substrate to the overall Raman
field enhancement depending on the desired application area. Fi-
signal for their spectrally separated bands even at low Ge content
nally, owing to high density of nanoparticles in the middle areas
(ωSi−Si =517.3 cm−1 at x=0.05).
(see Fig. 1d), one can find pairs of the nanoresonators having
nanometer-scale separations (gaps). Such dimer-like nanostruc-
Increase of the pump laser intensity I introduces the thermal-
tures are known to provide substantial EM field enhancement in
induced blueshift of the Si-Si band as it is illustrated by series
the gap area, which could further boost the biosensing perfor-
of intensity-resolved Raman spectra obtained from the isolated
mance of the proposed substrate. The calculation results provid-
230-nm diameter Si0.8 Ge0.2 nanoparticle (Fig. 3b). The mea-
ing EM field enhancement in Si1−x Gex dimer structures are given
sured spectral shift ∆ of this Raman band can be use to recalcu-
in Fig. S3 of the Supporting information.
late the average temperature increase ∆T of the nanoresonator
Composition tunable light-to-heat conversion of Si1−x Gex having a certain size D and chemical composition x according
alloyed nanoparticles. We further confirm the composition tun- to theoretical predictions from 16 . This function is provided in
able light-to-heat conversion in Si1−x Gex alloyed Mie nanores- Fig. S4 of the Supporting Information. By systematically mea-
onators by pumping the isolated nanostructures with a 532-nm suring the thermal-induced spectral shift of the Si-Si band for
laser radiation (see Methods for details), while simultaneously different alloyed Si1−x Gex nanoresonators (Fig. 3c), we experi-
detecting the spectral position and thermal-induced spectral shift mentally confirm the dependence of the light-to-heat conversion
of the Stokes Raman bands associated to the optical phonons. efficiency on both the relative Ge concentration x and nanoparti-

4| J
our
nal
Name,
[yea
r][
,vol
.
],
1–7
 Page 5 of 8 Nanoscale
View Article Online
DOI: 10.1039/C9NR01837A

cle base diameter D. Specifically, the alloyed nanoparticles having

higher Ge concentration (x=0.22) generally can be heated more
efficiently, which is in agreement with our numerical analysis
(Fig. 2a). Taking into account almost linear dependence of ∆T(I)
at high temperatures, the majority of Ge-rich nanoparticles in the
middle area of the sample is expected to reach irreversible mod-
ification threshold via melting (∆T≈1230 K) at I<28 mW/µm2 ,
which is confirmed using detailed high-resolution SEM inspec-
tion of the laser pumped nanoparticles (see Fig. S5 in Support-
ing information). To the contrary, alloyed nanoparticles having
low Ge content (x=0.07) show substantially lower heating rate
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

with an estimated threshold pump intensity of I<70 mW/µm2 Fig. 4 Temperature-feedback SERS detection of 4-ATP molecules.
(∆T≈1300 K), which is out of the range of our pump laser. Gener- SERS spectra of the 4-ATP molecules adsorbed on the isolated
ally, the observed difference in light-to-heat conversion efficiency Si0.93 Ge0.07 Mie resonator measured at pump intensities of 5.5 and

Nanoscale Accepted Manuscript

10 mW/µm2 . The characteristic bands of the 4-ATP at 1078 and
for nanoparticles having different relative Ge concentration x cor-
1587 cm−1 are highlighted by the green-color areas. The initial molar
relates well with our numerical analysis (Fig. 2a). concentration of the 4-ATP molecules in the alcoholic solution is 10−4 M.
Noteworthy, for the fixed Ge content x, alloyed nanoparticles Inset provides magnified section of these Raman spectra near the Si-Si
also demonstrate size-dependent heating efficiency. Specifically, band of nanoparticle.
for both probed Ge concentrations, Si1−x Gex resonators with the
diameter D around 220±15 nm show the most efficient heat-
ing definitely related to the EM field enhancement mediated by the nanoresonator-molecule system in the process of SERS mea-
MQ resonance excitation. Also, we found that in the range surements which is impossible to realize with plasmonic nanos-
of nanoparticle diameters available for our sample (D=200– tructures. For the alloyed Si1−x Gex nanoresonators with a fixed
300 nm), the light-to-heat conversion efficiency decrease with D Ge composition x, one can easily calibrate the appropriate pump
(Fig. 3c). Both these findings correlate well with our numerical intensity I according to the data from Fig. 3c to exclude the
calculations of the EM field structure inside the alloyed nanopar- temperature-induced destruction of the analyte. For the chosen
ticles (Fig. 2b). 4-ATP molecules with their critical temperature of ≈450 K mea-
Finally, the Si-Si Raman band of the alloyed nanoparticles sured using the thermogravimetric analysis (see Fig. S6 in the
is spectrally shifted with respect to the substrate mode at Supporting Information), the pump intensity I up to ≈6 mW/µm2
520.7 cm−1 even at room temperature ∆T=0 K (low pump in- can be used for safety non-destructive SERS measurements with
tensity). In this respect, the Raman signal coming from the Si the majority of alloyed nanoparticles with low Ge content.
substrate can be easily excluded from the overall collected in- Figure 4 shows the representative Stokes SERS spectra of
tensity providing high-contrast mapping of local position, density the 4-ATP molecules adsorbed on the isolated Si0.93 Ge0.07 Mie
and distribution of the nanoresonators on the substrate at sub- resonator measured at pump intensities of 5.5 mW/µm2 and
wavelength lateral resolution, for example, simultaneously with 10 mW/µm2 and signal accumulation time of 10 s. At such pump
any biosensing SERS experiment. Fig. 3(d,e) shows the opti- intensities, the Mie resonator can be heated either below or above
cal and SEM images of the representative section of the middle the thermal degradation temperature of the 4-ATP molecules as
area having the most dense arrangement of the nanoparticles. As one can verify by tracing the position of the Si-Si Raman band
seen, no separated nanoparticles can be resolved in the optical of the nanoparticle (see inset in Fig. 4). At both intensity lev-
image obtained with a dry 100× microscope lens with NA=0.9 els, main νCS(a1) and νCC(a1) vibration modes of the 4-ATP
(Fig. 3d). To the contrary, Raman mapping of the Si-Si band molecules at 1078 and 1587 cm−1 (marked by green area in
intensity (within the spectral range of 502–507 cm−1 ) allows Fig. 4) can be identified; however, at lower pump intensity we
to identify and resolve all 26 Si0.8 Ge0.2 nanoresonators located found twice higher SERS yield from the 4-ATP molecules. These
within the rather small 2 × 2 µm2 surface area (Fig. 3f). can be related to the fast thermal degradation of the analyte at
Temperature-feedback non-invasive SERS identification of high-intense laser irradiation, which is also seen from the in-
analyte molecules. Analysis of the calculated and measured creased luminescence background observed in the SERS spectra
optical properties of alloyed Si1−x Gex Mie resonators indicates at I=10 mW/µm2 . At lower pump intensity, SERS yield from ana-
that for detection of biological analyte molecules with generally lyte molecules remains stable even after 3 minutes of laser irradi-
low thermal degradation temperatures, utilization of nanoparti- ation, while at higher intensity level the characteristic signal dis-
cles with low Ge composition (x=0.07) supporting lower heat- appears after 30 s of irradiation. In particular, this shows a high
ing rate and larger outside EM field enhancement is the most stability of SERS measurements performed with careful temper-
appropriate choice. Symmetric Lorentz-shape profile of the Si- ature control. Finally, no Raman bands related to the dimercap-
Si band observed for all probed nanoparticles even at rather high toazobenzene molecules are observed during all Raman measure-
pump intensity indicates the uniform temperature distribution in ments on the Si1−x Gex nanoparticle indicating the weak chem-
the nanoresonator 18,19 . This allows to precisely define (with an ical invasiveness of the used substrate, which does not support
accuracy of about ≈10 K, 18 ) and control the local temperature in the light- or temperature-driven catalytic oxidation of the probed

J
our
nal
Name,
[yea
r][
,vol
.
], 1–7 | 5
 Nanoscale Page 6 of 8
View Article Online
DOI: 10.1039/C9NR01837A

molecules previously reported for plasmonic 39,40 and even metal- parency with the low current ion beam.
oxide substrates 14 . Raman microspectroscopy. A commercial Raman apparatus
(Alpha, WiTec) equipped with a pump CW laser (central wave-
3 Conclusions
length of 532 nm) is used to check the chemical composition of
We have shown that emission and light-to-heat conversion prop- the fabricated Si1−x Gex as well as to study light-to-heat conver-
erties of Si1−x Gex fully alloyed nanoresonators can be gradually sion in the nanostructures and their chemo- and biosensing per-
tailored via tuning relative Ge concentration. As a result of such formance. To do this, linearly-polarized pump radiation is focused
doping, position of the enhanced Si-Si Raman band has been de- by a 100× dry lens (Carl Zeiss) with a numerical aperture (NA)
tuned from that of bulk silicon substrate to enable background- of 0.9, producing the pump area size of ≈ 0.071 µm2 , which
free optical imaging of the Si1−x Gex nanoparticles at high-spatial is slightly larger than the average lateral size of the irradiated
resolution. We also highlighted importance of temperature con- Si1−x Gex nanoparticles (≈ 0.045 µm2 ). Focal depth of the pump
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

trol of the laser-pumped nanoantenna-analyte system in the pro- beam is estimated to be ≈ λ (NA−2 ) = 0.5 µm, which is high
cess of SERS measurements which can be realized in a facile way enough for uniform excitation of the nanoparticles. The pump
with crystalline Si1−x Gex alloyed nanoresonators. In this paper, intensity is varied within the range of 0.1 to 65 mW/µm2 using

Nanoscale Accepted Manuscript

we used commercial SOI substrate with a fixed thickness of Si top a built-in calibrated attenuator. The Stokes Raman signal col-
layer resulting in a certain size of the produced Si1−x Gex nanopar- lected from the focal volume with the same lens is analyzed using
ticles. However, one can easily tailor the average size of the al- a grating-type spectrometer (1800 lines/mm) equipped with an
loyed resonant nanostructures by changing the thickness of the thermo-electrically-cooled charge-coupled-device camera.
dewetted Si layer as it was previously reported in numerous stud- To access the biosensing performance of the Si1−x Gex nanopar-
ies 28,41–43 . Demonstrated versatile easy-to-fabricate biosensing ticles, we use the 4-aminothiophenol (4-ATP) as a model ab-
platform can find application for truly non-invasive optical iden- sorbate. These analyte molecules are known to support catalytic
tification of biomolecules, pollutants, explosives at low concentra- transformation to dimercaptoazobenzene and can be used as an
tions as well as tracing chemical reaction, providing an informa- indicator of chemical invasiveness of the SERS substrate. The sub-
tion on analyte‘s composition, analyte-nanoantenna temperature strate containing nanostructures is maintained in alcoholic solu-
response and their spatial position. tion of the 4-ATP at its molar concentration of 10−4 M for several
4 Methods hours and then rinsed in distilled water. The obtained point spec-
tra are averaged over at least 50 similar measurements from the
Fabrication of Si1−x Gex Mie resonators. We produce the similar nanostructures to perform statistical averaging.
Si1−x Gex nanoparticles via solid state dewetting of a silicon-on-
Numerical calculations. The distribution of the electric field
insulator substrate (with a 12-nm thick top Si layer and a 30-
amplitude E is obtained by numerical solving the wave Maxwell
nm thick SiO2 spacer) in UHV-MBE at 10−10 Torr and constant
equation using finite element method realized in commercial soft-
temperature of 800o C for 2 hours according to the method de-
ware package (Comsol Multiphysics). A linearly-polarized radia-
veloped in 29 . The typical background pressure in UHV-MBE is
tion has the electric field amplitude E0 and wavelength of 532 nm.
10−11 Torr, while the difference is caused by the presence of the
Dielectric constants for Si, SiO2 and Ge are taken from 35,44–46 .
evaporated Ge atoms. Under such conditions, the top Si layer is
The geometry of the nanoparticles is directly extracted from the
partially dewetted along the crystallographic orientations result-
SEM/TEM images. We have assumed that the Si and Ge compo-
ing in formation of crystalline faceted Si nanoparticles (central
nents inside the alloyed particles do not interact with each other.
area in Fig. 1b). Then, by maintaining the same temperature
Then, complex permittivity εr of the Si1−x Gex alloy can be ex-
we deposed 20 Ge monolayers above the sample to exchange the
tracted from the Clausius-Mossotti formula 47 :
matter composition on the surface resulting in formation of ran-
 
domly arranged Si1−x Gex nanoparticles under further dewetting εr − 1 4π
= N1 αGe + N2 αSi ,
process (middle and side areas in Fig. 1b). Noteworthy, in the εr + 2 3
central area the Ge deposition proceeds above the already formed
where Nx , the number of particles in a unit volume, is recalculated
faceted Si nanoparticles resulting in Ge-shell Si-core nanoparticle
according to the volume concentration of Ge, and polarizabilities
geometry. Alternatively, in the middle and side areas the fully al-
α are converted from their corresponding refractive indices.
loyed Si1−x Gex nanoparticles of hemi-spherical shape are formed
with the relative Ge concentration x=0.07÷0.22, which depends 5 Acknowledgements
on the distance from the central area (see Fig. S1 in Supporting
information for details). Experimental part of this work is supported by the Russian Sci-
Si1−x Gex Mie resonator characterization. Fabricated ence Foundation (Grant no. 18-79-10091). Numerical calcula-
Si1−x Gex nanoparticles are characterized using high-resolution tions are supported by Russian Foundation for Basic Researches
scanning electron microscopy (SEM, Carl Zeiss Ultra 55+) and (Grant no. 17-02-00538). M.A. acknowledges the LMON and
transmission electron microscopy (TEM). Specifically, for TEM IM2NP Laboratories for technical assistance with TEM studies.
study of the fabricated the nanoparticles, cross-sectional lamellas
References
are produced using focused ion beam lithography device (FIB, FEI
HELIOS 600). Lamellas are carefully thinned to electron trans- 1 I. Alessandri and J. Lombardi, Chem. Rev., 2016, 116, 14921.

6| J
our
nal
Name,
[yea
r][
,vol
.
],
1–7
 Page 7 of 8 Nanoscale
View Article Online
DOI: 10.1039/C9NR01837A

2 A. Krasnok, M. Caldarola, N. Bonod and A. Alú, Adv. Opt.

Mater., 2018, 6, 1701094. 23 J. Cambiasso, M. KoÌĹnig, E. CorteÌAs, ˛ S. SchluÌĹcker and
S. A. Maier, ACS Photonics, 2018, 5, 1546–1557.
3 A. Kuznetsov, A. Miroshnichenko, M. Brongersma, Y. Kivshar
and B. Luk’yanchuk, Science, 2016, 354, aag2472. 24 X. Li, H. Hu, D. Li, Z. Shen, Q. Xiong, S. Li and H. J. Fan, ACS
Appl. Mater. Interfaces, 2012, 4, 2180–2185.
4 P. Anger, P. Bharadwaj and L. Novotny, Phys. Rev. Lett., 2006,
96, 113002. 25 X. Li, G. Chen, L. Yang, Z. Jin and J. Liu, Adv. Funct. Mater.,
2010, 20, 2815–2824.
5 S. Cong, Y. Yuan, Z. Chen, J. Hou, M. Yang, Y. Su, Y. Zhang,
L. Li, Q. Li, F. Geng et al., Nat. Commun., 2015, 6, 7800. 26 G. Sinha, L. E. Depero and I. Alessandri, ACS Appl. Mater.
Interfaces, 2011, 3, 2557–2563.
6 R. Regmi, J. Berthelot, P. M. Winkler, M. Mivelle, J. Proust,
F. Bedu, I. Ozerov, T. Begou, J. Lumeau, H. Rigneault et al., 27 A. Balcytis, Y. Nishijima, S. Krishnamoorthy, A. Kuchmizhak,
P. R. Stoddart, R. Petruskevicius and S. Juodkazis, Adv. Opt.
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

Nano Lett., 2016, 16, 5143–5151.

Mater., 2018, 6, 1800292.
7 N. Bontempi, L. Carletti, C. De Angelis and I. Alessandri,
Nanoscale, 2016, 8, 3226–3231. 28 J. Ye, D. Zuev and S. Makarov, Int. Mat. Rev., 2018, 1–39.
8 M. Sanz-Paz, C. Ernandes, J. U. Esparza, G. W. Burr, N. F. 29 M. Naffouti, T. David, A. Benkouider, L. Favre, M. Cabie,

Nanoscale Accepted Manuscript

van Hulst, A. Maitre, L. Aigouy, T. Gacoin, N. Bonod, M. F. A. Ronda, I. Berbezier and M. Abbarchi, Nanotechnology,
Garcia-Parajo et al., Nano Lett., 2018, 18, 3481–3487. 2016, 27, 305602.
9 M. V. Zyuzin, D. G. Baranov, A. Escudero, I. Chakraborty, 30 M. Naffouti, T. David, A. Benkouider, L. Favre, A. Delobbe,
A. Tsypkin, E. V. Ushakova, F. Kraus, W. J. Parak and S. V. A. Ronda, I. Berbezier and M. Abbarchi, Small, 2016, 12,
Makarov, Sci. Rep., 2018, 8, 6107. 6115–6123.
10 N. Bontempi, K. Chong, H. Orton, I. Staude, D.-Y. Choi, 31 M. Naffouti, T. David, A. Benkouider, L. Favre, A. Ronda,
I. Alessandri, Y. Kivshar and D. Neshev, Nanoscale, 2017, 9, I. Berbezier, S. Bidault, N. Bonod and M. Abbarchi, Nanoscale,
4972–4980. 2016, 8, 2844–2849.
11 A. Vaskin, S. Mashhadi, M. Steinert, K. E. Chong, D. Keene, 32 M. Abbarchi, M. Naffouti, B. Vial, A. Benkouider, L. Lermusi-
S. Nanz, A. Abass, E. Rusak, D.-Y. Choi, I. Fernandez-Corbaton aux, L. Favre, A. Ronda, S. Bidault, I. Berbezier and N. Bonod,
et al., Nano Lett., 2019, 19, 1015–1022. ACS Nano, 2014, 8, 11181–11190.
12 A. Tittl, A. Leitis, M. Liu, F. Yesilkoy, D.-Y. Choi, D. N. Neshev, 33 N. Bontempi, I. Vassalini, S. Danesi, M. Ferroni, P. Colombi
Y. S. Kivshar and H. Altug, Science, 2018, 360, 1105–1109. and I. Alessandri, J. Phys. Chem. Lett., 2018, 9, 2127–2132.
13 Y. Gao, N. Gao, H. Li, X. Yuan, Q. Wang, S. Cheng and J. Liu, 34 S. Danesi, M. Gandolfi, L. Carletti, N. Bontempi, C. De Ange-
Nanoscale, 2018, 10, 15788–15792. lis, F. Banfi and I. Alessandri, Phys. Chem. Chem. Phys., 2018,
20, 15307–15315.
14 E. Mitsai, A. Kuchmizhak, E. Pustovalov, A. Sergeev, A. Miro-
nenko, S. Bratskaya, D. Linklater, A. Balcytis, E. Ivanova and 35 Handbook of Optical Constants of Solids, ed. E. D. Palik, New
S. Juodkazis, Nanoscale, 2018, 10, 9780–9787. York: Academic Press, 1998.
15 P. Dmitriev, D. Baranov, V. Milichko, S. Makarov, I. Mukhin, 36 E. R. Johnson and S. M. Christian, Phys. Rev., 1954, 95, 560.
A. Samusev, A. Krasnok, P. Belov and Y. Kivshar, Nanoscale, 37 We do not show this range in the Fig. 2a as it is out of the range
2016, 8, 9721–9726. of Ge concentrations x for the available sample.
16 M. Balkanski, R. Wallis and E. Haro, Phys. Rev. B, 1983, 28, 38 M. I. Alonso and K. Winer, Phys. Rev. B, 1989, 39, 10056.
1928. 39 M. Sun and H. Xu, Small, 2012, 8, 2777–2786.
17 H. Burke and I. Herman, Phys. Rev. B, 1993, 48, 15016. 40 J. Ye, J. Hutchison, H. Uji-i, J. Hofkens, L. Lagae, G. Maes,
18 G. P. Zograf, M. I. Petrov, D. A. Zuev, P. A. Dmitriev, V. A. G. Borghs and P. V. Dorpe, Nanoscale, 2012, 4, 1606–1611.
Milichko, S. V. Makarov and P. A. Belov, Nano Lett., 2017, 17, 41 F. Leroy, F. Cheynis, Y. Almadori, S. Curiotto, M. Trautmann,
2945–2952. J. Barbe, P. Müller et al., Surface Science Reports, 2016, 71,
19 M. Aouassa, E. Mitsai, S. Syubaev, D. Pavlov, A. Zhizhchenko, 391–409.
I. Jadli, L. Hassayoun, G. Zograf, S. Makarov and A. Kuch- 42 E. Bussmann, F. Cheynis, F. Leroy, P. Müller and O. Pierre-
mizhak, Appl. Phys. Lett., 2017, 111, 243103. Louis, New Journal of Physics, 2011, 13, 043017.
20 V. Milichko, D. Zuev, D. Baranov, G. Zograf, K. Volodina, 43 M. Aouassa, I. Berbezier, L. Favre, A. Ronda, M. Bollani,
A. Krasilin, I. Mukhin, P. Dmitriev, V. Vinogradov, S. Makarov R. Sordan, A. Delobbe and P. Sudraud, Applied Physics Letters,
and P. Belov, Laser Photonics Rev., 2017, 12, 1700227. 2012, 101, 013117.
21 R. M. Bakker, D. Permyakov, Y. F. Yu, D. Markovich, 44 M. A. Green, Sol. Energy Mater Sol. Cells, 2008, 92, 1305–
R. Paniagua-Domínguez, L. Gonzaga, A. Samusev, Y. Kivshar, 1310.
B. Lukyanchuk and A. I. Kuznetsov, Nano Lett., 2015, 15, 45 L. Gao, F. Lemarchand and M. Lequime, Opt. Express, 2012,
2137–2142. 20, 15734–15751.
22 M. Caldarola, P. Albella, E. Cortés, M. Rahmani, T. Roschuk, 46 G. E. Jellison Jr, Opt. Mater., 1992, 1, 151–160.
G. Grinblat, R. F. Oulton, A. V. Bragas and S. A. Maier, Nat. 47 C. Bottcher, Elsevier Publishing Company, 1952, 279.
Commun., 2015, 6, 7915.

J
our
nal
Name,
[yea
r][
,vol
.
], 1–7 | 7
 Nanoscale Page 8 of 8

View Article Online

DOI: 10.1039/C9NR01837A
Published on 01 April 2019. Downloaded by KEAN UNIVERSITY on 4/2/2019 9:33:59 PM.

Nanoscale Accepted Manuscript

Incorporation of Ge to Si1-xGex nanoparticles allows to modify their light-to-heat conversion and optical
response, being crucial for biosensing applications.