Ultrastrong Plasmon–Exciton Coupling by Dynamic Molecular

Sep 29, 2017 - Plasmon–exciton polaritons arise from the coherent coupling of the localized plasmon of metal nanoparticles and the exciton of nearby...
3 downloads 18 Views 4MB Size
Subscriber access provided by La Trobe University Library

Article

Ultrastrong Plasmon-Exciton Coupling by Dynamic Molecular Aggregation Francesco Todisco, Milena De Giorgi, Marco Esposito, Luisa De Marco, Alessandra Zizzari, Monica Bianco, Lorenzo Dominici, Dario Ballarini, Valentina Arima, Giuseppe Gigli, and Daniele Sanvitto ACS Photonics, Just Accepted Manuscript • DOI: 10.1021/acsphotonics.7b00554 • Publication Date (Web): 29 Sep 2017 Downloaded from http://pubs.acs.org on September 30, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Photonics is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

Ultrastrong Plasmon-Exciton Coupling by Dynamic Molecular Aggregation Francesco Todisco,† Milena De Giorgi,∗,† Marco Esposito,†,‡ Luisa De Marco,† Alessandra Zizzari,†,‡ Monica Bianco,† Lorenzo Dominici,† Dario Ballarini,† Valentina Arima,† Giuseppe Gigli,†,‡ and Daniele Sanvitto†,¶ †CNR NANOTEC - Institute of Nanotechnology, Via Monteroni, 73100 Lecce, Italy ‡Dipartimento di Matematica e Fisica ”Ennio De Giorgi” - Strada Provinciale Lecce-Monteroni, Universit´a del Salento, Campus Ecotekne, Lecce 73100, Italy ¶INFN - Istituto Nazionale di Fisica Nucleare, Sezione di Lecce - Via Monteroni, 73100 Lecce, Italy E-mail: [email protected]

Abstract Plasmon-exciton polaritons arise from the coherent coupling of the localized plasmon of metal nanoparticles and the exciton of nearby resonant nanoemitters. The behaviour of such systems is strictly defined by the initial choice of the metallic and excitonic materials, with only weak control possibilities, essentially limited to polarization related effects or photoswitchable molecules. Here we propose a new strategy to control the plasmon-exciton splitting, based on the number of excitonic dipoles involved in the interaction. By integrating plasmonic arrays in a microfluidic device and injecting a dilute near-infrared cyanine dye solution, we are able to probe in real time the emergence and evolution of the strong plasmon-exciton coupling regime. When dye molecules selectively aggregate on silver as a result of chemical affinity, we observe

1 ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a continuous increase of the Rabi splitting up to an exciton energy fraction as high as 35%, compatible with an ultrastrong coupling regime.

Keywords Plasmon, Plexciton, Ultrastrong Coupling, Plasmon Exciton Polariton, Plexciton Dynamics, Microfluidics, Strong Coupling Localized surface plasmon resonances (LSPR) in metal nanoparticles have attracted a great interest due to their unique optical properties, including the confinement of light in mode volumes far below the diffraction limit (V  λ3 ) and the strong local electromagnetic (EM) field enhancement. 1 These properties, together with the broadband room-temperature operation and extreme EM field spatial and spectral profile tunability, make LSPRs ideal candidates for a wide range of applications, including surface enhanced Raman spectroscopy (SERS), 2 plasmonic tags and particle-based therapies, 3 sensing, 4 energy harvesting, 5 and photochemical reactions. 6 The main drawbacks of plasmonic resonances still remain the intrinsic absorption losses of metals, that result in poor quality factor (Q) resonances, 7 and the extremely weak nonlinearities which essentially inhibit the direct active control of the LSPR optical response. 8 A promising strategy to overcome these limitations includes the coupling of metallic nanostructures to the resonant exciton of dipole emitters, thus creating hybrid plasmon-exciton mixtures in which the interaction between the participating modes enables to significantly modify their properties. In particular, when the involved modes are near-resonant and their coupling strength is stronger than losses, the system enters the so-called strong coupling regime, and new light-matter hybrid modes emerge, known as plasmon-exciton polaritons (PEPs, or plexcitons). 8 These modes, that manifest as a peak splitting in the transmission and scattering spectra of the system, are characterized by a typical anti-crossing behaviour in the energy dispersion,

2 ACS Paragon Plus Environment

Page 2 of 24

Page 3 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

with minimum energy separation (at zero detuning) called Rabi splitting. As hybrid systems, PEPs result in properties intermediate to those of the bare uncoupled states. So far interesting effects have been reported in these systems, such as spatial coherence in the exciton luminescence, 9 magneto-optical induced properties in non-magnetic molecules, 10 chemical tailoring of molecular states 11 and lasing. 12 Strong coupling between excitons and LSPRs has been extensively studied with gold, 13 silver, 14 and, more recently, aluminum 8,15 nanostructures, either with individual or array of nanoparticles. Although gold is considered a good and stable plasmonic material, it is limited to work at energies below 2.4 eV where the onset of interband transitions damps the LSPR. The same happens for aluminum at around 1.5 eV. On the contrary, the onset of interband losses in silver is shifted at higher energies (∼3.1 eV) allowing the realization of low losses devices with resonances in the visible spectral range. 7 Unfortunately, this metal reacts very quickly with gases composing the air, resulting in oxidation and sulfidation processes 16 that sensibly degradate the LSPR spectrum. On the other hand, as far as the active materials are concerned, the high oscillator strength and sharp excitonic resonances of quantum dots, 17 molecular dyes 18 and j-aggregates 19 have been often used in PEPs systems. Up to now, the efforts to actively control PEPs have been focused on the use of photocromic molecules 20 and polarization controlled plasmonic resonances, 13 resulting in a switchable Rabi splitting. An effective dynamic control of the strong plasmon-exciton coupling has only been observed with microcavities and surface plasmon polaritons coupled with photoswitchable molecules, 21 preventing the exploitation of plexcitonic nanoantennas for possible sensing applications and effective devices. 22 To make a step forward in this context, we have integrated plasmonic nanoantennas in a microfluidic chamber (MFC), in order to study in real time the build up of plasmon-exciton strong coupling. We clearly observe an anti-crossing behaviour of the PEP dispersion, and a continuous increase of the Rabi splitting due to the increasing aggregation of molecules at the surface of the metallic nanostructures. In particular, for sufficiently long immersion times, the number of deposited molecules par-

3 ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ticipating to the coupling becomes high enough to reach the regime of ultrastrong coupling.

Sample Architecture A sketch of our sample is depicted in Figure 1. Eight 70×70 silver nanodisk (ND) square arrays are defined onto a glass substrate by electron beam lithography followed by thermal evaporation. The NDs are 40 nm thick, and have diameter (D) ranging from 50 nm to 190 nm, and center to center interparticle distance three times the ND diameter, so as to minimize near-field interactions between neighbor nanoparticles while maintaining a constant NDs filling factor. The nanostructured sample is integrated in a microfluidic device, realized using photolithography and wet etching methods, as detailed in the Methods section.

Figure 1: Schematic illustration of the sample architecture: several square arrays of silver nanodisks with different diameter are vertically aligned (top left) and integrated in a microfluidic device (bottom). A scanning electron microscope image of the as fabricated 190 nm diameter silver nanodisks is shown in the top right panel. Scale bar is 1 µm.

The ND arrays are optically characterized on an upright microscope coupled to a spectrometer equipped with a CCD camera. Broadband white light from a tungsten lamp is focused on the sample surface by a condenser lens with low numerical aperture ( 0

0.6

0.4

0.2

0

D=150 nm Detuning ~ 0

Extinction (1-T)

0.8 0.6 0.4 0.2 0

D=170 nm Detuning < 0

0.8

Extinction (1-T)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

Extinction (1-T)

Page 9 of 24

0.6

0.4

0.2

0 400

500

600

700

800

900

Wavelength (nm)

Figure 4: Extinction spectra at different time delays, for nanodisks with (a) 50 nm, (b) 110 nm, (c) 150 nm and (d) 170 nm diameter, corresponding to different plasmon-exciton detunings. The black solid lines indicate the corresponding bare plasmon extinction spectra, measured in the 2-propanol filled MFC, while the black dashed line indicates the extinction from uncoupled liquid and solid dye. The arrows indicate the evolution of the extinction peaks. 9 ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

organic polariton systems, 12,33 in our case the lower branch is much more visible than the upper one, which is reabsorbed from higher energy molecular states. On the other hand, the continuous increase of the splitting with increasing delay, clearly visible near zero detuning, indicates a corresponding increase of the coupling strength as more and more dye molecules aggregate on the NDs surface. However, at longer delays (∆t > 50 ) the absorption from uncoupled molecules rapidly degradates the splitting visibility, as shown in the Supporting Informations.

Coupled Oscillators’ Analysis A quantitative indication of the interaction strength between plasmon and exciton, can be obtained by the effective coupling strength given by 2~ΩR /~ωexc , where ~ωexc is the exciton energy, and 2~ΩR is the Rabi energy splitting. It is well known that the latter is proportional q Nf , where N is the number of dipoles coupled to the electromagnetic field, f is the to V dipole oscillator strength, and V is the optical modal volume. In the particular case of strong coupling with broad excitonic resonances, like those measured here, Houdr´e et al. have demonstrated 34 that the Rabi splitting is independent on the homogeneous or inhomogeneous linewidth of the exciton, and it only depends on its energy-integrated absorption



α(E)dE, according to the Smakula equation for defects in

a host medium. 28,35 It follows that, despite the presence of more than one excitonic peak in the solid dye absorption spectrum, it is possible to describe the entire system with a single exciton effective frequency (ωX ), linewidth (γX ), and oscillator strength (fX ). We can thus determine, without loss of generality, the effective plasmon-exciton coupling strength by fitting the experimental data with a two coupled oscillators’ model, using the full Hopfield hamiltonian

H = ωpl a† a + ωX b† b + ΩR (a† + a)(b† + b) + Ω2R /ωX (a† + a)2 {z } | {z } | Hint

H0

10 ACS Paragon Plus Environment

(1)

Page 11 of 24

b

a

c Δt≈1'

Δt=0

190 nm

Δt≈5'

170 nm

Extinction (a.u.)

150 nm

130 nm

110 nm

90 nm

70 nm

50 nm

400

500

600

700

800

900 400

Wavelength (nm)

d

Peak Positioh (nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

500

600

700

800

900

400

Wavelength (nm)

e 900

900

800

800

800

700

700

700

600

600

600

500

500

500

60

80

100

120

140 160

Diameter (nm)

180

400

60

80

100

120

140 160

Diameter (nm)

180

400

600

700

800

900

Wavelength (nm)

f

900

400

500

60

80

100

120

140 160

Diameter (nm)

180

Figure 5: Extinction spectra of the nanodisk arrays, and relative peaks position measured at different delays after the microfluidic chamber filling: (a,d) ∆t=0, (b,e) ∆t ≈1’ and (c,f) ∆t ≈5’. Black arrows in the spectra indicate the peaks position. The solid black line in the diameter-dependent dispersions indicates the bare plasmon dispersion, while the dashed lines indicate the fitting of the extracted peaks position through a two coupled-oscillators model, given by Equation 2.

11 ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 24

where H0 is the basic Hamiltonian describing self-energies of the exciton and the plasmon, and Hint represents the interaction terms. a† (a) and b† (b) are the plasmon and exciton creation (annihilation) operators, respectively, whereas ωpl is the plasmon frequency. Note that quantum formalism is used here for simplicity but same results are obtained with a semi-classical approach. The exact solution of Equation 1 leads to the following dispersion equation:

ω 4 − ω 2 ((ωpl − ıγpl )2 + (ωX − ıγX )2 + g 2 ) + (ωpl − ıγpl )2 (ωX − ıγX )2 = 0

(2)

where γpl is the plasmon linewidth, and g is the coupling strength. Equation 2, once fixed the plasmon and exciton parameters, enables the fitting of the polariton dispersion at a given coupling strength. In particular, it is demonstrated that the peaks splitting only depends on the energy detuning, while the bare modes linewidth only influences the polariton linewidth, with the particular property that, for sufficiently high interaction energy, only the homogeneous linewidth of the coupled modes matters. 34 In typical plasmon-exciton systems, the plasmon lifetime (≈ 10 fs) is much shorter than the molecular one (≈ 1 ns), and the polariton linewidth is thus essentially determined by its plasmonic part, as defined by the Hopfield coefficients. 36 However, when the inhomogeneous broadening becomes relevant, as in our case, it comes into play in determining the polariton linewidth 34 and an upper limit for an effective exciton linewidth γX can be extracted from the polariton linewidth. In this way we find γX = 80 meV, as described in the Supporting Informations. For the plasmon counterpart, the frequency and linewidth, ωpl and γpl , are evaluated by fitting the extinction spectra in Figure 2b, finding a nearly constant linewidth, γpl ≈ 110 nm, and the LSPR dispersion shown as a red line in Figure 2c. On the other hand, the effective exciton frequency is left as a free parameter for the fitting of the experimental anticrossing. Doing this we find λX = 815 nm and the PEP dispersions shown as dashed lines in Figure 5e-f. This value is significantly different from that of the diluted dye solution (775 nm) 12 ACS Paragon Plus Environment

Page 13 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

while it is closer to the lowest energy band of the molecules in solid phase (828 nm), as a further demonstration that the strong coupling regime arises between the plasmon and the solid aggregated dye. The evidence that the coupling with this resonance prevails on the higher energy one of the solid state dye (centered at 725 nm) can be an indication that j-type aggregates with high oscillator strength are formed at longer wavelengths. 37 The coupling strengths g(∆t), extracted from the dispersions’ fitting, are reported in Figure 6. Here, a clear increase of the coupling as a function of the delay appears, with two main features. On the one hand the lowest splitting, measured at one minute delay, is characterized by a coupling strength of ~ΩR =178±27 meV. This value, considering that q 2 γpl γ2 + 2X = 144 meV, γpl ≈ 200 meV near zero detuning, satisfies the inequality ~ΩR > 2 thus confirming that in our system the coupling is much larger than losses, and the strong coupling regime is accessible. 38 On the other hand the coupling strength increases up to 275±42 meV after about five minutes. Given that the plasmonic modal volume, the losses and the molecular oscillator strength are fixed, this coupling increase can only be correlated to an increase of the number of molecules N aggregated on the NDs surface and participating √ to the coupling. As a consequence, given that ΩR ∝ N , we can estimate a relative increase ∆N/N ≈ 140% of the number of molecules with respect to the first splitting observed. Interestingly, for ∆t ≥ 3 minutes, the number of interacting molecules results in a Rabi splitting that is a significative fraction of the exciton energy, ~ΩR > 30%EX , as shown in Figure 6. Under this circumstance the system evolves from the strong to the so-called ultrastrong coupling regime, where new interesting effects have been predicted to arise, including squeezed vacuum and generation of correlated photons. 39,40

Number of Dipoles’ Evaluation An estimation of the absolute number of molecules involved in the coupling can be obtained by making some considerations on the molecular transition dipole moment (µ) and the

13 ACS Paragon Plus Environment

ACS Photonics

40 30

0.2

Strong

0.1

Weak

20 10

� R ��� X� �%�

Ultrastrong

0.3 g (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 24

0

0 0

1

2 3 4 Time (min)

5

Figure 6: Plasmon-exciton coupling strength at different delay from the filling of the microfluidic chamber. The error bars, evaluated from the fitting, increase with time as a consequence of the broadening of the molecule absorption peak. plasmonic mode volume (V ) of the silver NDs. The coupling strength, in fact, can be written as 41 s gN =

4π~N cµ2 λ0 V

(3)

where N is the number of molecular dipoles, λ is the exciton wavelength and  is the environment dielectric constant. The molecular transition dipole moment can be directely related to the oscillator strength f using 42

f=

8π 2 me λµ2 e2 3ch

(4)

where me and e are the electron mass and charge, and c is the light velocity in vacuum. By using Equation 4, and considering f = 1.4, as reported in literature for a similar molecule, 30 we obtain µ = 5.1 × 10−29 C · m = 15.4 D. We can use this value to calculate the number of coupled molecules N in Equation 3, once the environment dielectric constant  and the plasmonic mode volume V are defined. The former, to a first approximation, can be assumed as that of the 2-propanol surrounding the nanostructures,  = 1.88 at λ = 800 nm. 43 On the other hand, the exact calculation of the mode volume is an unsettled problem for leaky cavities, including plasmonic nanoan14 ACS Paragon Plus Environment

Page 15 of 24

tennas. However, an estimation of the modal volume has been reported for spherical silver nanoparticles, based on the divergence normalization when integrating the electromagnetic field over spherical domains with different radii. 44 Here modal volumes of around 2.5 × 10−4 µm3 are calculated for a sphere of 50 nm radius, that can be considered as an indication of the modal volume of our cylindrical nanoparticles. We can thus calculate the coupling strength in Equation 3 as a function of both the mode volume V and the number of participating molecules N . The result is shown as a color map in Figure 7. Here, the thresholds for the weak and the ultrastrong coupling regimes are reported as dashed lines, according to the system losses and the fraction of coupling strength and exciton energy, respectively. As expected from Equation 3, a linear dependence of V on N exists for a fixed coupling strength. In particular, from the coupling extracted in our system, indicated as black lines in Figure 7, we deduce that the number of molecules participating to the coupling is of the order of few thousands, as indicated by the dots, their exact value

m

eV

me

=2 47 ,g

' ≈5

Δt

5 27

m

eV

=

250

5

200

3

1

300

Coupling Strength, g (meV)

Ul Stro me tra n V str g C o n ou g C pli ou ng Δt pli ≈4 ng ', g

78

g= 19 9

g=1 1',

350

2',

Δt≈

7

Δt≈

Weak 9 Coupling

V

remaining highly dependent on the exact estimation of the modal volume.

Mode Volume, V (x10-4�m3)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

150 5 10 15 Number of Molecules, N (x103)

20

Figure 7: Coupling strength in colour scale, calculated as a function of the plasmonic modal volume V and of the number of molecules involved in the coupling N . The experimental values of the coupling are indicated as solid lines, together with the thresholds for the onset of the strong and ultrastrong coupling regimes (dashed lines). The horizontal line indicates the modal volume of a silver sphere with 50 nm radius, as calculated in ref. 44.

15 ACS Paragon Plus Environment

ACS Photonics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Conclusions We have experimentally demonstrated the possibility to observe the dynamic evolution of plasmon-exciton interaction, from weak to strong and ultrastrong coupling regime. This possibility is offered by the proper combination, in a microfluidic device, of silver nanostructure arrays and a near infrared cyanine dye, characterized by a large coupling strength and a strong affinity with silver. In such a system, the gradual and selective aggregation of molecules on the nanostructures’ surface, allows to continuously increase the number of dipoles participating to the coupling, and to overcome the threshold for the onset of the ultrastrong coupling regime. Moreover, we have shown that strong coupling can arise even using broad linewidth excitonic materials if the number of the involved dipoles and their coupling strength is large enough to overcome losses. We believe that our system can be an optimum candidate for the study and control of light-matter interaction in properly engineered hybrid plasmonic systems.

Experimental Methods Nanofabrication Silver nanodisk arrays are defined on a glass substrate by electron beam lithography. The substrate is first cleaned in acetone and 2-propanol. Then a 250 nm poly(methyl methacrylate) (PMMA) layer is spin coated at 6000 rpm and soft-baked at 180◦ C for 3 min. A 2 nm thick chrome layer is thermally evaporated to prevent charge effects in the electron beam writing procedure. The arrays are written by a Raith 150 system at 22 pA beam current and 30 keV. After electron exposure, the Cr layer is completely removed by a ceric ammonium nitrate based wet etching for 20 s and rinsed in water. The exposed resist is developed in MIBK:IPA solution in a 1:3 ratio for 3 min and rinsed in 2-propanol for 1 min. After thermal evaporation of 40 nm of silver, a liftoff process was performed in an mr-Rem 500 remover

16 ACS Paragon Plus Environment

Page 16 of 24

Page 17 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Photonics

solution (Microresist Technology) and rinsed in 2-propanol.

Microfluidic device fabrication A B-270 glass, accurately cleaned with acetone and 2-propanol, is patterned by photolithography: 45 a square shape chamber with a 5 mm side and two channels with nominal width of 1 mm are transferred from a photo-mask to the photoresist. After the geometry transfer, the glass substrate is etched in order to obtain a chamber depth of 3 mm with a buffered oxide etch solution by using a microwave reactor system. 46 The total internal volume of the chamber is of about 0.1 µL. Two holes of 1/32” diameter are processed into microchannels using a microdriller and diamond-coated drill bits, thus allowing to connect tubings for fluids introduction and removal. Finally, the etched B-270 glass and the patterned substrate are faced and sealed by means of a hot-melt gasket.

Optical characterization The arrays are optically characterized in transmission configuration. White light from a tungsten lamp is focused on the sample with an adjustable numerical aperture condenser lens, used with NA