Letter pubs.acs.org/NanoLett
Precise Attoliter Temperature Control of Nanopore Sensors Using a Nanoplasmonic Bullseye Colin R. Crick,† Pablo Albella,‡ Binghao Ng,‡ Aleksandar P. Ivanov,† Tyler Roschuk,‡ Michael P. Cecchini,† Fernando Bresme,† Stefan A. Maier,*,‡ and Joshua B. Edel*,† †
Department of Chemistry and ‡Department of Physics, Imperial College London, South Kensington Campus, London, SW7 2AZ, United Kingdom S Supporting Information *
ABSTRACT: Targeted temperature control in nanopores is greatly important in further understanding biological molecules. Such control would extend the range of examinable molecules and facilitate advanced analysis, including the characterization of temperature-dependent molecule conformations. The work presented within details well-defined plasmonic gold bullseye and silicon nitride nanopore membranes. The bullseye nanoantennae are designed and optimized using simulations and theoretical calculations for interaction with 632.8 nm laser light. Laser heating was monitored experimentally through nanopore conductance measurements. The precise heating of nanopores is demonstrated while minimizing the accumulation of heat in the surrounding membrane material. KEYWORDS: Nanopore, plasmonics, nanoplasmonics, field enhancement, temperature control, metallic nanopore
T
The increased heat caused by the presence of the plasmonic structures results in observed changes in molecular properties (in this case, the removal of charge cations at elevated temperatures).14 Accurate temperature control could be used for a range of purposes, namely improving the solubility of analytes at elevated temperatures, variation of translocation speeds, and probing the stability of analyte conformation.15−17 Optimization of nanopore temperature control requires focusing on two main areas: improving the heat absorption mechanism and thus increasing attainable temperatures, and also specifically targeting the nanopore and not the surrounding membrane.18−20 Nonspecific heating of nanopores has been reported in the literature, carried-out via the use of tightly focused lasers. These systems have demonstrated scalable heating of nanopores; however, the efficiency of these mechanisms is relatively low as there is no direct absorption of incoming radiation.21 The selective absorption of a particular wavelength of radiation includes examples of using infrared absorbing dyes and also thin metal films, which act to directly convert laser energy into heat.22,23 The use of plasmonic structures has been shown to be capable of selectively absorbing radiation and has demonstrated highly efficient energy absorption and subsequent heating of the immediate environment (attoliter volumes).24 Plasmonic structures used for targeted heating include nanoparticle/
he development of sensors for the detection and analysis of biological analytes has seen a recent rise in academic and commercial research interest due to their vast applicability in the monitoring of a wide range of biological systems.1−4 In particular, nanopores, representing a subset of biosensors utilizing nanometer-sized holes for the label-free detection of single molecules, hold immense promise in becoming highly efficient single molecule detectors that are sensitive enough to fully characterize analytes as they travel (translocate) through the nanopore.3,4 Advances in nanopore fabrication techniques and greater understanding of analyte translocations may soon lead to in situ analysis of molecular properties that would only require a single molecule.5,6 The design of new nanopore geometries and the incorporation of innovative features are crucial in progressing toward this goal. To achieve this goal, ionic current blockade sensing has been the most popular.7 More recent alternative detection strategies used to complement this include the use of tunnelling currents, fluorescence, and surface-enhanced Raman spectroscopy in an effort to increase sensitivity.8−12 Localized environmental conditions within, and in close proximity to, nanopores are extremely important when considering the detection of any analyte. Variation of environmental factors could have a range of effects on an analyte.4,13 For instance, direct heating of the detection volume of the nanopore system could lead to a change of local nanopore conditions and could extend the applicability of a nanopore system. An example of this includes the use of gold nanoparticles as antennae for absorbing incoming radiation. © XXXX American Chemical Society
Received: October 17, 2014 Revised: November 27, 2014
A
dx.doi.org/10.1021/nl504536j | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
plasmonic feature arrays.1,25−27 These examples show the importance of periodic features in maximizing the electromagnetic field enhancement through supporting plasmonic resonance.28 Bullseye structures, constructed from plasmonically active materials, provide both the repetition of features to provide plasmonic enhancement and have central symmetry and can be made to act as a focal point.29 Variations in the material and geometrical parameters of the plasmonic structure allows for the tuning of the wavelength at which maximum absorption of radiation occurs.30−32 This grants extra control of the local environment of the nanopore for advanced bioanalyses. Existing approaches in combining plasmonic structures with nanopores suffer from several disadvantages, such as polarization dependence and the inability to selectively target the nanopore using a specified excitation wavelength.14,18 Ideally, one would like a plasmonic nanopore sensor that allows for control of the local environment of the pore, not just via the power of the excitation light, but which is polarization independent and tunable via a single geometrical parameter. In this paper, we demonstrate how solid-state nanopores surrounded by a bullseye structure enable efficient nanopore heating using a monochromatic 632.8 nm light source. The heating will be targeted specifically at the nanopore, limiting the absorbance of the membrane. In our device this volume is 1 aL; however, more typically, it would be in the zL to aL range.13 In particular, a plasmonic bullseye structure milled into a gold film on top of a free-standing membrane of SiNx is used to demonstrate this. The specifications of the plasmonic nanopore design are first justified by the presentation of simulated data. Devices were fabricated and characterized from optimized geometries obtained through electromagnetic modeling. The temperature change within the nanopore was determined using ionic conductivity measurements and compared with predicted results. The ionic conductivity provides a direct indication of the environment within the nanopore, and the measurements are highly sensitive and reproducible. The temperature changes within the system presented are rapid (both heating and cooling); this is due to the tight focus of the incident laser and quick absorbance of light. The efficiency, accuracy, and practicality of nanopore heating is then discussed. Figure 1a shows the bullseye structure used to guide electromagnetic (EM) energy toward the groove. The purpose of using the bullseye structure is to drastically enhance the EM fields within the nanopore via surface plasmon polaritons (SPPs), which are excited when phase-matching conditions are met by the rings. The ring period which is crucial for the SPP phase-matching is estimated to be 518 nm for an excitation wavelength of 632.8 nm at normal incidence and using the SPP dispersion relation given by33 β=
εmεd εm + εd
Figure 1. Device schematic. (A) Illustration showing a top-down view of features milled into gold film; (i) ring period 518 nm, (ii) pore diameter 80 nm, and (iii) ring width 80 nm. (B) Illustration showing a side-on cross-section of freestanding membrane; (i) gold layer with plasmonic structures and a nanopore of 100 nm and (ii) silicon nitride membrane with a nanopore of 100 nm. (C) Top-down view, SEM image showing the nanopore/bullseye structure milled into the membrane (scale bar shows 1 μm). (D) Complete experimental setup. The freestanding membrane (60 × 60 μm) with bullseye structure (Ø = 4 μm) is irradiated with laser light (on the gold side), submerged in 1 M KCl, and a potential is applied across the membrane.
of the free-standing SiNx/Au membrane were 60 μm × 60 μm. The bullseye structure was then milled into the gold side of the membrane using a focused ion beam (FIB, Ga+) (see Figure 1A).10 Figure 1B shows the cross section of our nanopore. As can be seen, only the center pore cuts through the entire membrane, the rings of the bullseye structure exist only in the gold film. The pore size, ring period, and ring width were 80, 518, and 80 nm, respectively. Note that the ring period corresponds to the excitation of SPPs on the bullseye structure at an excitation wavelength of 632.8 nm, as mentioned before. The entire nanopore assembly was simulated using COMSOL Multiphysics 4.3a. The incident radiation is xpolarized and normally incident. The symmetry of the system causes the plasmonic nanopore to be polarization independent.36 The field enhancement, defined as |E|/E0, in the nanopore is monitored and plotted in Figure 2. As can be seen, there is indeed a broad peak in field enhancement around 625 nm in the case where the bullseye structure, with 518 nm periodicity, is present (see blue solid line). This is in stark contrast to the case where there is no bullseye structure, which shows a field enhancement of only one to two times around 625 nm. Figure 2b−d shows the field distribution at 532, 632.8, and 685 nm, respectively; these are the laser wavelengths used in the experimental work (profile views are included in the Supporting Information, S2). Again, we see that there is strong field enhancements corresponding to the excitation of SPPs at 632.8 nm (|E|/E0 ∼ 7), while at the other two wavelengths, the field enhancements are comparatively weak (|E|/E0 ∼ 2). The effect of ring periodicity on the SPP excitation efficiency has been explored by modeling a bullseye system with 250 nm periodicity (Supporting Information, S1). As expected, this has
(1)
where β is the propagation constant of the SPP and εm and εd are the permittivities of the metal and dielectric, respectively. In our case, the metal was assumed to be gold and modeled using permittivity data from Johnson and Christy, while the dielectric was assumed to be water with a refractive index of 1.33.34 A full explanation of this is provided in the Supporting Information (S1). Free-standing membranes, consisting of 100 nm Au on top of 100 nm SiNx, supported by 300 μm thick silicon wafers were fabricated using standard photolithography.35 The dimensions B
dx.doi.org/10.1021/nl504536j | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
film to act as an EM energy absorber, as we shall see later in this paper.37 Additionally, factors such as the surface area of the surface terminations of gold (for, e.g., the edges of the rings) in contact with the environment will affect the way metallic loss translate into heat. The full optimization of such a complex system is out of the scope of this paper. Here, we focus on how the excitation of SPPs can result in a more efficient conversion of EM energy to temperature changes as compared to cases where there are no SPPs excited. The experimental temperature changes are calculated from ionic current measurements (Figure 3A); however, theoretical temperature changes are modeled directly, these are shown in Figure 3B/C alongside experimental results. The simulated temperature change spectra (Figure 3C) use the temperature at the center point of the nanopore, that which is in the same plane as the external gold surface. The temperatures are modeled for the bullseye structure (blue solid line) and that of just the nanopore with no rings (black-dashed line). The initial temperature T0 was 25 °C, and the intensity of the incident light was 20 mW/μm2, corresponding to the experiments. As can be seen (Figure 3), the temperature change in the bullseye case is higher than that without rings. Notably we can see a slight peak in temperature change at around 632.8 nm for the bullseye, confirming that expectations that the excitation of SPPs leads to a larger temperature change. Nanopore heating experiments were carried out on a custom built optical microscope (for details see Supporting Information).10 The entire free-standing membrane was mounted between two reservoirs filled with an electrolyte solution (1 M, KCl, 10 mM Tis 1 mM EDTA pH 8) (Figure 1D). Ag/AgCl electrodes were placed in each reservoir for the measurement of ionic current across the nanopore. The pores were heated using 532, 632.8, and 685 nm excitation sources incident from the Au side of the membrane. This choice of wavelengths allowed the probing the optimal wavelength for pore field enhancement and the energies in proximity to this. For each device, multiple chronoamperometric traces were measured (ranging from −300 mV to 300 mV with a 25 mV step) while varying the excitation sources and power. IV curves were extracted from the chronoamperometric traces. The inset in Figure 3A shows representative normalized I−t traces for 200 mV. Importantly, even at maximum excitation source power, the ionic current showed excellent baseline stability over time, with a relatively small increase of the ionic current noise. For the data set presented in Figure 3A the ionic current noise increased from 0.2 nA (RMS at 200 mV) for a control measurement without a laser to 0.4−0.8 nA at maximum excitation source power. The ionic current data was then used to estimate the corresponding temperature change. The ionic conductivity of the nanopores system is dependent on the temperature within the detection volume.18,21 This is the most constricted volume of the system, i.e., the immediate volume of the nanopore. The temperature variations within this detection volume will directly affect the current allowed to flow at a given potential. Irradiation of the nanopore with the lasers (532, 632.8, and 685 nm) and the subsequent temperature rise are accompanied by a rise in the ionic conductivity of the pore. As the pore temperature cannot be easily measured directly, the rise in pore conductance and ionic current were used to estimate the temperature change. The fundamental relationship between the temperature of a solution and ionic current can be established by directly measuring these changes in bulk solution; this has been previously reported in the literature.18
Figure 2. (A) Normalized field enhancement within the nanopore, for case of no plasmonic bullseye (slab, black), the 518 nm periodicity bullseye (blue), and the 250 nm periodicity bullseye (gray). (B) Field distribution at 532, 632.8, and 685 nm.
the effect of shifting the broad peak in EM enhancement to around 685 nm (Figure 2A, gray), which is the region required for phase matching with 685 nm light. Propagation of SPP on the gold−water interface is thus achieved, and a maximization of the field enhancement at (|E|/E0 ∼ 7) occurs. Field distribution images for the 250 nm periodicity bullseye devices are shown in the Supporting Information (S3). It is important to mention that the main source of heat originates from the absorption (resistive losses) of light. This is directly related to the near-field enhancement in the pore; however, other factors affecting the actual heat generation and diffusion also contribute. This would include physical aspects of the device such as taking into account the entire surface area of ring structures, which when varied can also impact the source of heating. Another factor is the volume of gold used for modeling, which would affect the diffusion of heat away from the pore. A general observation of this is that, when using a constant illumination spot size, a larger bullseye period will provide a larger amount of heating. We would expect the high field enhancement to be a result of increased charge oscillations within the metal. At the same time, the charge oscillations would result in metallic losses that are propagated throughout the system as heat, causing the temperature to rise. Thus, from the analysis of the field enhancements within the hole, we can begin to see how the excitation of SPPs on the bullseye structure might lead to increased temperature changes in the nanopore as compared to the case when there are no SPPs excited. Of course, the bullseye nanopore system proposed here is quite complicated, and such a simplistic picture breaks down as we approach the interband transition wavelength of gold causing the entire gold C
dx.doi.org/10.1021/nl504536j | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 3. Temperature change estimated from current change upon laser irradiation of nanopore with and without plasmonic bullseye structure. (A) Experimental I−V plots used to calculate temperature changes at the nanopore. Representative plots are shown for irradiation of nanopores with surrounding plasmonic bullseye, using the 532, 632.8, and 685 nm lasers at power of 5.58, 5.64, and 5 mW, respectively. The I−V plots were extracted from multiple I−t traces measured at potentials ranging from −300 mV to 300 mV with a 25 mV step (Inset). Representative I−t traces using aforementioned laser powers at a single potential (200 mV), the traces were normalized for direct comparison of the ionic current noise at no laser and full laser power. (B) Laser power dependence of temperature change using the three different wavelength lasers. (C) A plot of the wavelength dependence of temperature change. For (B/C), experimental results are shown by the points plotted with error bars, whereas simulated results are indicated with lines. (D) Simulated temperature maps for both slab and bullseye conditions at the various laser wavelengths (Profile views are available in the Supporting Information S4).
to approaches based on assembled Au nanopaticles.14 For example, using a single nanoparticle temperature increases of 41.1 °C were achieved for 150 mW excitation, which approximates to ∼30× higher laser power. The results demonstrate good agreement with the predicted results (Figure 3), with lower radiation wavelengths giving a greater efficiency of heating. An enhancement of pore heating with bullseye structure was observed for all laser wavelengths, relative to a single pore. The amount of enhancement did however vary. It was shown that the 532 nm laser provided the largest estimated temperature enhancement increasing when irradiated with 5 mW of laser power. This is a result of the greater efficiency of light absorption at this wavelength, facilitated by interband transition of gold in the membrane. As a result of this, the 532 nm laser also caused the heating of the entire area irradiated local to the nanopore (shown in Figure 3D), in effect heating the pore externally. The 685 nm lasers show the lowest temperature rise, caused by providing the lowest induced field enhancement and little absorption by
Figure 3 shows the conductivity changes upon irradiation with the different lasers; this was carried out for a range of potentials. The corresponding conversion from ionic current to estimated temperature change is expressed below:18 ΔT = (a /b + T0) × ΔI /I0
(2)
where a/b are experimentally established constants, T0 is the ambient solution temperature, ΔI is the change in measured current upon irradiation, and I0 is the current flow through the nanopore without laser irradiation. A temperature increase of 47 °C is achieved when a nanopore, surrounded with plasmonic bullseye, is irradiated with the 632.8 nm laser at power of 5 mW (Figure 3C). This is higher than the 25 °C temperature rise for just the pore. The enhancement in heating is quite substantial, given that both the modeled and experimental temperature change is +140 and +88%, respectively, with the presence of a bullseye structure. The temperature increase also occurs linearly with laser power (Figure 3B), in agreement with eq 2. The heating demonstrated is significantly more efficient compared D
dx.doi.org/10.1021/nl504536j | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
linearly with temperature difference is Δφ = SΔT,39 hence following the same dependence we observe in the experimental data. The magnitude of the emf potential is determined by the Seebeck coefficient, S. The Seebeck coefficient varies widely across electrolytes, being in the μV/K range for alkali halides.40 At infinite dilution the Seebeck coefficient of KCl can be extracted from heat of transport estimates, ∼50 μV/K at 25 °C.41 Temperature differences of the order of 100 °C could then produce an emf of about 5 mV, which are small as compared with the applied potential. This contribution can be smaller if we account for a reduction of the Seebeck effect with concentration. The importance of these thermal gradient induces effects was addressed by computing the transient change of temperature with time using the COMSOL Multiphysics package. Figure 4
the gold membrane. The 632.8 nm laser was shown to give substantial heating, which was limited to the local area of the pore (Figures 2B/3D). The area of concentrated heating is limited to within the rings of the bullseye (Supporting Information, S5), with the strongest heating confined to within the smallest ring. The area of strongest heating can be estimated to be
