1. Introduction

Being an active scanning probe microscopy (SPM), tip-enhanced Raman spectroscopy (TERS), which roots in advanced scanning instruments such as atomic force microscope (AFM) [1] and scanning tunneling microscope (STM) [2], not only be able to gain strong Raman signal intensity and extremely high spatial resolution but also can map 2D or 3D Raman images across sample surface [3, 4]. Based on these detection advantages, TERS has made a great successful improvement in development of Raman devices to date. However, two physical principles of nanosize tapered metal probes (including metal-coated and metal nanoparticle-attached probes) exploited in TERS devices limit the TERS to be a more flexible and universal instrument. First, enhanced local electric field generated by surface plasmons at apex of a needle-like metal probe decays rapidly with increasing distance away from the tip [5]. The probe has to be very close to a detected sample (<5nm in most of cases) to obtain high enhancement factor and spatial resolution in mapping Raman spectroscopy. The interaction between the metal tip and the sample induced by the tiny space, such as electrostatic, van der Waals, and hydration forces, disturbs molecular resonances and deduces shift of intrinsic Raman peaks at some level [6, 7]. This is a considerable disadvantage for the TERS to identify a target from a complex mixture mingling several types of similar molecules. Second, high dissipative loss of electromagnetic energy in metal material and small scattering cross-section of nanosize tapered probe greatly reduce collection efficiency in near field and cannot transfer the energy over long distance to far field [8]. Therefore, a commonly used TERS system exploiting a tapered metal probe has to use a low-efficient far-field conventionally optical lens to collect near-field scattered photons [9]. The additionally adding lens makes the TERS be a semi-near-field Raman microscope with complex optical structure actually.

Limited by the two physical restrictions of metal probes, performances of TERS devices have to be carried out under one or more special conditions, such as extreme cold and high vacuum [2], low reliability and repeatability [10], single species of molecule. Therefore, some applications of Raman spectroscopy, particularly in nanobiology, need an ideal scanning near-field Raman microscopy (SNRM) device to avoid these performance limitations and provide a stable and reliable Raman spectrum by non-contact way under more naturally detecting conditions. This means a new type of probe is required to break the restrictions of metal probes and provide enhancement of excitation and collection within tens of nanometers detection range near the probe tip. To achieve this goal, some different types of probes, such as adiabatic plasmon nanofocusing probe [11], bowtie nanoaperture antenna probes [12], coaxial optical antenna scan probes [13], and campanile geometry probe [14] have been reported and discussed experimentally. Some of the probes can provide strong local electric field, nevertheless, no probe which has nanosize tapered pattern can efficiently collect energy within tens of nanometers range from the probe tip.

Fortunately, some previous works have demonstrated that optoplasmonic (OP) structures [15, 16], which are composed of dielectric structures (e.g. microspheres) and metal nanoparticles (NPs), not only can enhance electric field in the interval of two metal NPs but also can transmit emitted energy from the interval over long distance to far field efficiently. Herein, these OP components provide many potential applications in the field of nanocircuits [15], nanosensor [17], and surface enhanced Raman spectroscopy (SERS) detections [18, 19]. Prior applications of OP structures mainly base on high quality factor resonance mode and characteristics of electric field in the gap between metal NPs, however, the quasi-stable enhancement of electric field with wide range wavelength outside the gap in the vicinity of metal NPs has not been discussed yet.

In this paper, we propose a conceptual OP model and investigate electric field characteristics in the vicinity of the model tip. After electric field distributions of the model have been calculated numerically with wide wavelength range in near field, three attractive near-field characteristics can be concluded, including: 1, enhanced electric field within tens of nanometers distance leaving from the model tip; 2, localized electric field distributions at nanoscale in the near field; 3, efficient collection and delivery with broadband quasi-stabilization for emitted energy. Benefiting from these characteristics, the proposed OP model can be used as a scanning near-field probe to enhance Raman scattering and harvest scattered photons simultaneously within tens of nanometer distance away from the probe tip. Hence, the OP probe ablates the restrictions of metal probes and brings out a possibility of realizing an ideal SNRM for broadening application range of Raman spectroscopy in nanotechnology.

2. Near-field electric field characteristics of the OP probe

The proposed OP probe model is featured in Fig. 1. Surrounding by air condition in free space, a gold nanobowtie is placed on surface of a silica microsphere which has diameter of D. Two horizontally opposed triangular patches with interval d form the bowtie. The thickness and the side length of the triangular patch are h and l respectively. To avoid sharp corner of the triangles that can induce unreliable results in simulation, all angles of the triangles are rounded with radius r. The model is located in a Cartesian coordinate system, and the origin of the coordinate system is located at the geometrical center of the upper surface of the gold bowtie, shown at right in Fig. 1. To demonstrate a general physical concept rather than optimized engineering results, and meet the actual processing realization, the geometric parameters of the model are set to D = 3μm, d = 10nm, h = 40nm, l = 120nm, and r = 10nm respectively. Electric field distributions of the model were calculated using with finite-difference time-domain (FDTD) method. The optical constants of gold and silica are obtained from experimental data [20] and [21] respectively. Although magnetic enhancement in nano structures is a interesting field [18, 22], because enhancement of Raman signal is more sensitivity with electric field than with magnetic field [23], then the electric field is only concerned factor in this paper. According to these setup and method, the calculation results of three characteristics of the OP probe are described next in sequence.

 

Fig. 1 Schematic of the OP probe model combining a silica microsphere and a gold nanobowtie used in Raman detection.

Download Full Size | PDF

2.1 Near-field electric field enhancement

When an incident planar light (electric field vector is x-polarized) with wavelength $\lambda$ illuminates the OP probe from the bottom of the microsphere along the positive direction of z axis, then an enhanced electric field distribution is generated in the vicinity of the nanobowtie, especially near the faced apexes as shown in Fig. 1. To sketchily assess electric field near the probe, the $\lambda$ was firstly set from 600nm to 1050nm (since this band is more compatibility and less harm to biomolecular activity than shorter band). Then electric field enhancement factor $f_E$ of the probe can be valued with$f_E=E_m/E_0$, where $E_m$ and $E_0$ are electric fields produced with and without the probe structure in simulation respectively. Since the $E_0$ is the electric intensity of the incident light in air condition and calculated to 1, then the $f_E$ is equal with $E_m$. Herein, the enhancement of electric field is simplified with normalized intensity of electric field $\left|E\right|$ that produced by the probe.

Because the z axis is vertical geometric symmetry axis of the probe in the upward direction and represents center of the electric field distribution produced by the OP probe, we calculated the $\left|E\right|$ on the z axis from the origin to a few tens of nanometers distance to get the enhancement data. As shown in Fig. 2(a), the $\left|E\right|$ is draw as a function of parameters $\lambda$ and the Z, where Z represents the distance from the origin along the z axis. This calculation result shows two apparent features of the electric field enhancement. One is the electric field oscillates with changing $\lambda$ in the range of 600-1050nm, which mainly ascribes to the WGM effect of the microspheres [15, 16, 24, 25] The other is the electric field decays exponentially with increasing Z. This is accord with the physical principle of the surface plasmons, which contribute the strong electric field on the apexes. Although these two disadvantages affect stability and detecting range of the OP probe, the result still shows that the probe can provide a quasi-stable and strong electric field with long detection distance in wide wavelength range.

 

Fig. 2 (a) Normalized electric field $\left|E\right|$ yielded by the OP probe in broadband, Z represents the distance leaving from the origin along the z axis. (b) Illuminating with 785nm, $\left|E\right|$ yielded by the OP probe ($E_m$), individual microsphere ($E_{sphere}$), and individual nanobowtie ($E_{bowtie}$) on the z axis. The individual microsphere and individual nanobowtie are the two components of the probe.

Download Full Size | PDF

In order to clearly detail superiority of the electric field enhancement produced by the OP probe, $\lambda$ = 785nm that normally used in biomolecular detection was selected expressly to illuminate two contrasting models (individual gold nanobowtie and individual silica microsphere, resulting from that the microsphere and the nanobowtie were removed from the OP model respectively). Using same calculation parameters as previous described, electric fields of the two models on the z axis were calculated. Shown in Fig. 2(b), electric field results of the two models and the OP probe were dawn as a function of distance Z. This comparison intuitively demonstrates that the probe has stronger electric field within distance of 70nm than those of the individual nanobowtie and the individual microsphere. Concerning that the electric field of the microsphere ($E_{sphere}$) is almost stable and space between two declining curves, the electric field of the OP ($E_m$) and the electric field of the bowtie ($E_{bowtie}$), keeps approximate width with increasing Z, it is easy to find that the microsphere contributes this the strong $E_m$. The principle involving in this contribution is the nanojet effect on the silica microsphere [26], which gathers more incident energy onto surface of the microsphere to excite the plasmon resonance to produce strong localized electric field at the bowtie apexes [27].

Moreover, because the silica microsphere is an optical diffraction component, the $E_{sphere}$ can be employed as a standard line for dividing effective detection distance of Raman spectroscopy. Based on the standard line, the effective distances of the OP probe and the individual nanobowtie are >70nm and ~15nm respectively. Comparing the distance of the individual nanobowtie is longer than 5nm (which is the limitation in most TERS devices) and similar with the result of these works on SERS [28], the OP probe demonstrates obvious advantage in detection range for Raman detections. This result is also due to the nanojet of the microsphere. In addition, the nanobowtie can be regarded as two horizontally and oppositely placed tapered metal probe, this assumption is helpful to express the OP probe breaks the first restriction of metal probes from another aspect.

2.2 Localized electric field distributions in the near field

High spatial resolution is one of prominent advantages of TERS devices. To verify whether the OP probe can offer nanoscale spatial resolution, localized electric field distributions produced by the OP probe were investigated in the near field. Because monochromatic laser usually be used to excite Raman spectrum, we chose incident wavelength 785nm to illuminate the probe. The characteristics of localized electric field is evaluated by full width at half maximum (FWHM) of the electric field distribution near the probe along x and y directions on horizontal cross section (x-y plane) with distance Z. As shown in Fig. 3(a), the FWHMs on two coordinate axes are drawn as a function of Z in the x and y directions.

 

Fig. 3 (a) FWHM of electric field distributions in the x direction (solid line) and y direction (dash line) on the x-y plane, Z is distance between the x-y plane and upper surface of the nanobowtie of the OP probe. (b) Electric field distribution in x-y plane when Z = 0nm, where means the upper surface. The maximal electric field is 981.7 at hot spots, while the electric filed at the center is 381.9. (c) Electric field distribution in x-y plane when Z = 10nm. The maximal electric filed is 115.7 at the center.

Download Full Size | PDF

In the y direction, we can notice that there is an approximately linear relationship between the Z and the FWHM. The relationship can be expressed with FWHM = 18.8 × Z + 19.8. According this, we can find that the FWHM rises with increasing Z, the approximate linear factor is 18.8. This result indicates that the probe can provide a nanoscale resolution (<100nm) in the range of Z<40nm in the y direction.

In the x direction, the result of FWHM shows that an extreme high spatial resolution can be achieved down to sub-nanometer when a sample is very close to the bowtie surface (Z$\to$0). This is due to two 'hot spots', where produce highest electron density and strongest electric field, at the faced vertices of the opposite triangles. Figure 3(b) shows the electric field distribution on the upper surface of the bowtie (Z = 0nm) and indicates the two hot spots and the maximal electric filed is 981.7 (this is bigger than the electric field 381.9 shown in Fig. 2(b) at middle point between the two hot spots owing to decay). However, for the decay of the electric field with increasing Z, the FWHM decreases rapidly to ~20nm when the Z moves to ~3nm. This result is similar with the principle of conventional TERS devices, which provide high spatial solution and strong Raman signal depending on tiny space between probe apex and samples. When the Z moves further to ~25nm, the FWHM ascends slightly faster than that in the y direction. After that, the FWHM tends to be stable relatively and less than that in the y direction. Comparing to the approximate linear FWHM in the y direction, this variation in the x direction clear represents that the hot spots is main factor for deciding the FWHM.

Although the FWHMs in both x and y directions increase with Z, the probe still breaks the diffraction limit of 785nm light and provides nanoscale FWHM within 60nm distance from the probe tip. As an example shown in Fig. 3(c), the electric field distribution on the x-y plane is a regular and parabolic shape when Z = 10nm. This localized distribution is different from that in Fig. 3(b). In Fig. 3(c), the FWHM in the x and the y directions are 49nm and 39nm respectively, and the strongest electric field is 115.7, which just locate at the geometric symmetry center in the plane. This example demonstrates that the probe can provide regular localized electric field distribution over tens of nanometers and hence offer nanoscale spatial resolution in the near field for trapping and detecting biomolecules.

2.3 Near-field collection enhancement

The ability of OP structure that can transfer energy from the gap of metal nano dimers to far field has been reported [15]. In order to examine ability of the OP probe that can harvest emitted energy efficiently from tens of nanometers leaving away from the gap of the metal nanobowtie and transfer the near-field energy over long distance to far field, the incident planar light in Fig. 1 is substituted with an electric dipole, as shown in Fig. 4(a). The dipole is placed on the z axis spacing distance $d_d$ from the upper surface of the gold bowtie, and the electric field vector of the dipole is along the x direction with wavelength$\lambda$. When $d_d$ = 10nm and $\lambda$ = 785nm, electric field distributions in vertical cross section (x-z plane) with and without the OP structure are shown in Fig. 4(b) and Fig. 4(c) respectively. Comparing the two figures, the enhancement of energy collection and transmission of the probe is very intuitive.

 

Fig. 4 (a) Schematic of the OP probe excited by an electric dipole. The rectangular transmission plane is used to calculate the far field energy. (b) and (c) are electric field distributions in the x-z plane with the probe and without the probe respectively, when $d_d$ = 10nm and wavelength of the dipole is 785nm.

Download Full Size | PDF

In order to further demonstrate the enhancement of the OP probe by numerical values, a square transmission plane with length L = 3μm is horizontally placed near the bottom surface of the silica microsphere, as shown in Fig. 4(a). The transmission plane is used to calculate energy transmitted downward through it. The calculated results of the transfer energy with the OP probe and without the probe are expressed with ${\Gamma }_m$ and ${\Gamma }_0$ respectively. Then the enhancement factor $A$ of the probe is expressed with $A={\Gamma }_m/{\Gamma }_0$. As shown in Fig. 5, the $A$ is delineated as a function of $\lambda$ ranging from 625nm to 1025nm at four different $d_d$(0nm, 10nm, 20nm and 30nm). The result in Fig. 5 shows that bigger $A$ can be obtained when the OP probe is closer to the dipole. The strongest $A$ reaches 13455.9 when $d_d$ = 0nm at $\lambda$ = 786nm. But the $A$ decays rapidly to 1044.5 ($d_d$ = 10nm),183.1($d_d$ = 20nm) and 55.9 ($d_d$ = 30nm) with the same $\lambda$ = 786nm. This is just because the coupling efficiency of bowtie and dipole energy decreases with the increase of distance. Moreover, the $A$ shows an approximately same oscillating pattern within same $\lambda$ range, despite the difference of the $d_d$. This oscillation is mainly caused by the effect of WGM of the microsphere, just as the same as the oscillation of the electric field enhancement described above.

 

Fig. 5 Near-field collection enhancement $A$ of the OP probe in wide band with different distance $d_d$ between the dipole and the nanobowtie surface.

Download Full Size | PDF

The two characteristics of the $A$, the decay with the increasing distance and the oscillation with $\lambda$, are disadvantages for the energy collection and transfer efficiency, nevertheless, the OP probe still shows a better performance in harvesting and transferring near-field energy than that of conventional metal probes. For example, when the $d_d$ is 10nm, the maximum and the minimum of the $A$ are 1044.5 ($\lambda$ = 786nm) and 29.6 ($\lambda$ = 627nm) respectively, and the average value of the $A$ reaches 350.9. This result shows a significant advance of the OP probe in near-field collection. Moreover, if a detected biomolecular is excited by $\lambda$ = 785nm, Stokes scattering wavelength of a biomolecular sample should be larger than 785nm. This means that, within 30 nm detection range, the OP probe can provide a quasi-stable and continuous collection enhancement for Raman signal. The result may promote potential applications in field of weak and ultra fast laser detections to avoid heat damage of biological molecules.

3. Discussions

Because enhancements of the excitation and the collection are produced from collective effects of the two components, the nanobowtie and the microsphere, amending any morphology parameter of the two components, such as size and material, can change the wavelength response for incident and scattered light [29, 30]. If samples need another excitation wavelength being different from 785nm, a new OP probe model should be optimized. Benefiting from lots of works on the dielectric microsphere and metal bowtie structure [30, 31], it's easily to design an optimized OP probe with suitable wavelength response to satisfy different detected analyte. Moreover, it should be noted that because electromagnetic field coupling efficiency between regular bowtie structure and scattered photons is sensitivity with the direction of electric field vector of the photons. If electric field vector of the scattered photons is random and cannot induce plasmon resonance on metal tips of the gold bowtie in the probe model, the collection enhancement will be greatly counteracted. Therefore, the bowtie in this probe model would be replaced with other structures that are not sensitive to orientation of electric field vector, such as asymmetric [32] and symmetry breaking [33] structure. Although these peculiar nano structures are difficult to process, collection enhancement can be preserved in arbitrary detection environment.

In addition, comparing the results in Fig. 2(a) and Fig. 5, two similar features of the values of the $\left|E\right|$ and the $A$, decreasing with distance and oscillating with wavelength, can be noted. The oscillation origins from the WGM effect in the dielectric microsphere. Eliminating the WGM effect can help to make the quasi-stable be stable and obtain smooth enhancements of excitation and collection in wide wavelength range. Further, the two features demonstrate that the OP probe generally obeys the principle of optical reversibility near the vertical geometry centerline (the z axis) in the near field. The reversibility verifies that the microsphere in the OP structure is a reversible optical component undoubtedly and the nanobowtie shows optical symmetry along the z axis. This means that any horizontally opposed two metal NPs, if the NP has two symmetrical planar surfaces being normal to the z axis (such as cylinder, rectangle and polygon), can be exploited to form a OP probe with a dielectric micro lens, such as sphere and oval. Using physical and chemical fabrication methods, there are many ways to develop this type of OP probe for practical applications [34, 35].

4. Conclusions

The near field electric field distributions of the OP probe, which is combined of a silica microsphere and a gold nanobowtie, demonstrate three valuable characteristics. First, the probe can provide greater electric field intensity and longer effective detection range than those of a tapered metal probe. Second, the probe brings out nanoscale electric field distribution within distance of tens of nanometers. Third, the probe can efficiently gather energy in the near field and transfer the energy to far field with broadband, this is a significantly difference from any conventional nanosize probe. Benefiting from the three characteristics, the restrictions of traditional metal probes is dispelled for Raman applications. Exploiting this concept of the OP probe and works on the surface plasmons, advanced micro-nano fabrication techniques, and techniques in the SPM, an ideal near-field scanning Raman microscope is expected to be realized and to extend Raman applications greatly, especially in active nanobiology detections that require a natural, non-contact and non-destructive detection mode. More than that, the characteristics of the OP probe also offer opportunities in trapping techniques, high-resolution optical imaging, nano lithography and nano information devices to make a SNRM be a versatile and multipurpose instrument.