3D nanointerface enhanced optical microfiber for real-time detection and sizing of single nanoparticles.

Portable devices, which can detect and characterize the individual nanoparticles in real time, are of insignificant interest for early diagnosis, homeland security, semiconductor manufacturing and environmental monitoring. Optical microfibers present a good potential in this field, however, are restricted by the sensitivity limit. This study reports the development of a 3D plasmonic nanointerface, which is made of a Cu-BTC framework supporting Cu3-xP nanocrystals, enhancing the optical microfiber for real-time detection and sizing of single nanoparticles. The Cu3-xP nanocrystals are successfully embedded in the 3D Cu-BTC framework. The localized-surface plasmon resonance is tuned to coincide with the evanescent field of the optical microfiber. The 3D Cu-BTC framework, as the scaffold of nanocrystals, confines the local resonance field on the microfiber with three dimensions, at which the binding of target nanoparticles occurs. Based on the evanescent field confinement and surface enhancement by the nanointerface, the optical microfiber sensor overcomes its sensitivity limit, and enables the detection and sizing of the individual nanoparticles. The compact size and low optical power supply of the sensor confirm its suitability as a portable device for the real-time single-nanoparticle characterization, especially for the convenient evaluation of the ultrafine particles in the environment. This work opens up an approach to overcome the sensitivity limit of the optical microfibers, as long with stimulating the portable real-time single-nanoparticle detection and sizing.

Introduction

Portable devices, which can detect and characterize the individual nanoparticles in real time without the use of labels, have attracted significant demand for applications in early diagnosis, homeland sepan class="Chemical">curity, semiconductor manufacturing and environmental monitoring [1], [2], [3], [4]. The enhanced levels of ultrafine particulates in air, which is one of the major concerns for environment and health, may cause respiratory incidences and cardiac diseases [5]. It has been reported that the ultrafine particulate pollutants induce oxidative stress and mitochondrial damage [6] as well as cause the mutagenicity and cytotoxicity [7], along with altering the expression of genes in the human neurons [8]. Harmful nanoscale biological agents such as viruses (e.g. SARS, CoVID-19, AIDs, etc.) are also a threat to the human health, causing serious diseases or death [9]. Currently, a variety of nanoparticle detection and sizing methods are used, for example, transmission electron microscopy (TEM), scanning electron microscopy (SEM) and atomic force microscopy (AFM) [10]. However, these approaches tend to be cumbersome, expensive, and time-consuming. Therefore, a portable, cost-effective, less-environmentally demanding device with low-power supply, which can detect and size the individual nanoparticles in real time, is of vital need. Optical microfibers present good potential in this field, owing to their advantages of low cost, flexibility and micrometer-scale [11]. Nevertheless, the currently used optical microfibers are only suitable for measuring the collective behavior of the nanoparticles and molecules due to the relatively low sensitivity [12], [13], [14]. Overcoming the sensitivity limit to detect individual nanoparticles or molecules is a significant challenge [15]. Great efforts have been devoted to improve the sensitivity of the optical microfibers [16], [17], [18], [19], [20], however, it is still difficult to achieve the single-particle detection and even more difficult to attain the sizing.

The weak exterior evanescent field is the main issue strongly limits the detection sensitivity, since the refractive index (RI)-based optical microfibers detect the RI changes on the surface induced by the particle-binding interaction through their evanescent field. Thus, the tight confinement and pan class="Gene">surface enhancement of the evanescent field along the microfibers are beneficial for achieving high sensitivity with extremely low optical power [15]. The localized-surface plasmon resonances (LSPRs) in near infrared (NIR) range consistent with the evanescent field wavelength of microfibers, displaying the features of surface localization and near-field enhancement, can be expected to confine and enhance the evanescent field [21], [22], [23]. Compared to the novel metallic nanostructures, the semiconductor nanostructures, as an emerging family of plasmonic materials, are excepted to display LSPR effect in the NIR region [24], [25], [26]. The copper phosphide (Cu3-xP) nanocrystals, offering higher LSPR tunability in the NIR region and much better compatibility [27], [28], represent a promising alternative in this field. A 3D nanointerface containing plasmonic nanoparticles could further regulate the surface energy distribution on optical microfiber to meet the requirement of particle-binding detection. Metal organic frameworks (MOFs), due to the unique structural advantages [29], [30], are a kind of promising scaffold to support the plasmonic nanoparticles to form a 3D nanointerface. Herein, we develop a portable optical microfiber sensor functionalized by a plasmonic nanointerface made of the Cu3-xP nanocrystals supported by a 3D Cu-based metal–organic framework (Cu-BTC, BTC = 1,3,5-benzenetricarboxylic acid) (Fig. 1 A and 1B). The Cu3-xP nanocrystals are successfully embedded, with its LSPR tuning to coincide with the evanescent field of the optical microfiber. As the scaffold of nanocrystals, the 3D Cu-BTC framework assists the nanocrystals to confine the local resonance field on the microfiber surface with three dimensions, where the binding of the target nanoparticles occurs (Fig. 1C). Based on the evanescent field confinement and surface enhancement by the Cu-BTC supporting Cu3-xP nanocrystals (Cu3-xP @Cu-BTC) nanointerface, the optical microfiber overcomes its sensitivity limit and enables the detection and sizing of the single nanoparticles (Fig. 1D). A broad band source (BBS) provides a large emission bandwidth of 1250–1650 nm with low power supply, and an optical spectrum analyzer (OSA) is employed to monitor the transmission spectrum of the microfiber (Fig. 1A). The compact size along low power supply confirms the suitability as a portable device for real-time single nanoparticle detection on various occasions, especially for the convenient evaluation of the ultrafine particles in the environment. This work opens up the opportunities to overcome the sensitivity limit of the optical microfibers by constructing a 3D nanointerface, as well as presents a stimulus for the portable real-time single-nanoparticle characterization.

Fig. 1

A) Schematic of the optical setup. B) Photo of the optical microfiber sensor. C) Schematic of the Cu3-xP@Cu-BTC nanointerface on the sensing region. D) Schematic of the stepwise shift in the transmission spectrum induced by single-nanoparticle binding.

Experimental

Fabrication of Cu-BTC frameworks

The bulk Cu-BTC particles were dispersed in an N, N-dimethylformamide (DMF) solvent to form a suspension with concentration of 1 mg/ml. Then, the suspension was ultrasounded using an ultrasonic probe. After standing for 4 h, the sediment was removed. The supernatant was reserved for the next step.

Synthesis of Cu3-xP nanocrystals

The synthesis of Cu3-xP nanocrystals was carried out following the previous work [31] with minor modifications. The typical synthesis process please see the Experimental Details in the Supplementary data.

Fabrication of Cu3-xP@ Cu-BTC composites

The Cu3-xP nanoparticles were modified by amino groups and combined with the Cu-BTC fragments through the electrostatic attraction. The typical synthesis process please see the Experimental Details in the Supplementary data.

Interface functionalization of optical microfiber

The interface functionalization of the optical microfiber was carried out following our previous work [32] with minor modifications. The typical synthesis process please see the Experimental Details in the Supplementary data.

Results and discussion

Plasmonic nanocomposite design and tuning

In order to tune the LSPR frequency to the operational wavelengths of the optical microfiber, four groups of Cu3-xP nanocrystals labeled as Cu3-xP-I, Cu3-xP-II, Cu3-xP-III and Cu3-xP-IV were prepared as shown in Figs. S1 and 2A. For Group-I to Group-III, the nanocrystals presented LSPR frequencies from 1000 to 1250 nm (Fig. S1). The Group-IV nanocrystals had good crystallinity [27], and were observed to be hexagon nanoplates having a side length of 30 nm (Fig. 2 A and 2B). These nanocrystals presented an LSPR frequency at 1450 nm, which was consistent with the operational wavelength of the optical microfibers (Fig. 2C). Thus, Group-IV nanocrystals were employed in this work as the plasmonic nanoparticles.

Fig. 2

A) TEM image of the Cu3-xP nanoplates. B) High-resolution TEM (HRTEM) image of the Cu3-xP nanoplates, (inset: electron diffraction). C) Extinction spectra of the Cu3-xP nanoplates, Cu-BTC nanoparticles, and Cu3-xP@Cu-BTC nanoparticles. D) TEM image of the Cu-BTC nanoparticles. E) XRD patterns of the Cu-BTC fragments, Cu3-xP nanoplates and Cu3-xP@Cu-BTC nanoparticles. F) TEM image of the Cu3-xP@Cu-BTC nanoparticles. G) EDS mapping of the Cu3-xP@Cu-BTC nanoparticles. XPS spectra of the H) Cu-BTC nanoparticles, I) Cu3-xP nanoplates, and J) Cu3-xP@Cu-BTC nanoparticles.

The Cu-BTC framework was obtained by ultrasonicating the crystalline pan class="Chemical">Cu-BTC composite to get the supernatant fragments. The morphology of the Cu-BTC fragments is shown in Fig. 2D, which indicated that the Cu-BTC framework was irregular and thin-layered. The X-ray powder diffraction (XRD) pattern in Fig. 2E implies the existence of amorphous structures, which is proved to have an optimal packaging effect, thus, enabling the nanoparticles packaged on exhibit higher apparent activity [33]. The Cu3-xP nanocrystals were surface modified with amino groups and compounded with Cu-BTC through electrostatic attraction. The TEM image in Fig. 2F, SEM image in Fig. S2 and the energy dispersive X-ray spectroscopy (EDS) mapping in Fig. 2G indicate that the Cu3-xP nanocrystals were closely packed on the Cu-BTC surface. The XRD patterns of the Cu3-xP@Cu-BTC hybrids in Fig. 2E confirm the presence of the amorphous structure of the Cu-BTC fragments and crystalline structure of the Cu3-xP nanoplates. The chemical composition of the Cu-BTC fragments, Cu3-xP nanocrystals and the prepared composites demonstrated by the X-ray photoelectron spectroscopy (XPS) spectra in Fig. 2H, 2I, and 2 J also indicate that the Cu3-xP@Cu-BTC hybrids were composed of the elements Cu, O and P, which is in good compliance with the EDS element mapping analysis. The presence of N element, on the other hand implies the successful surface modification of Cu3-xP nanocrystals [34]. It confirms the successful formation of Cu3-xP@Cu-BTC hybrids. Notably, the compound on Cu-BTC fragments tuned the extinction peak of the Cu3-xP nanocrystals from about 1450 nm to 1550 nm (Fig. 2C), which was more compatible with the communication band and was expected to enhance the evanescent field more effectively. Thus, immobilizing on the Cu-BTC framework gave the plasmonic Cu3-xP nanocrystals a 3D structure, endowed it a higher apparent activity, a better resonance coupling and a resonance frequency more compatible with the evanescent field of microfiber.

3D interface on optical microfiber

The Cu3-xP@Cu-BTC composites were functionalized on the surface of a silica optical microfiber to form an interface, as shown in Fig. 3 A. The silica microfiber was fabricated from a single-mode fiber by heating and stretching processes [35]. The diameter of the microfiber was 6.4 µm, and the waist length was 3.9 mm (Fig. S3). The resulting structure functioned as an intermodal interferometer and generated interference fringe in the transmission spectrum, which resulted from the interference between the fundamental (HE11) and the higher-order (HE12) modes (Fig. S4). Positive charges were endowed on the silica microfiber surface through sequential modifications with piranha and 3-aminopropyl-triethoxysilane (APTES) solutions [36]. The Cu3-xP@Cu-BTC composites were functionalized on the modified microfiber surface to form an interface through electrostatic attraction [37]. The SEM images in Fig. 3B-3D visualize the evolution of the microfiber surface from the smooth silica surface (Fig. 3B) to sporadic decoration by Cu3-xP (Fig. 3C), followed by uniform decoration by the Cu3-xP@Cu-BTC composites (Fig. 3D). It is worth noting that the Cu-BTC fragments assisted the Cu3-xP nanocrystals to combine densely and evenly on the microfiber surface. By comparing Fig. 3D and Fig. S5, it can be obtained that the functionalization of bulk Cu-BTC materials was easy to cause local aggregation, resulting in the spectrum loss, while the Cu-BTC fragments made the interface thickness more even and avoid the spectrum loss (Fig. S6). Also the loading amount of Cu3-xP@Cu-BTC composites affected the quality of optical spectrum, so as to affect the detection ability of the sensor as shown in Fig. S7. The EDS mapping in Fig. 3E indicates the presence of C, Cu and P elements on the microfiber surface, thus, confirming the successful functionalization of Cu3-xP@Cu-BTC as a nanointerface. Thus, the Cu-BTC framework increased the quantity of Cu3-xP conctructing on the microfiber surface. And the plasmonic nanocrystals were constructed to a three-dimensional structure, which was more conducive to the binding of sensing targets.

Fig. 3

A) Sketch showing the Cu3-xP@Cu-BTC interface functionalized optical microfiber sensor. SEM images of the B) silica microfiber surface, C) microfiber surface with Cu3-xP functionalization and D) microfiber surface with Cu3-xP@Cu-BTC functionalization. E) EDS mapping of the microfiber with Cu3-xP@Cu-BTC functionalization.

Evanescent-field confinement and enhancement in three dimensions

Fig. 4 A shows the bulk RI sensitivity of the as-prepared sensor compared with that of the silica microfiber sensor without interface. The pan class="Chemical">silica microfiber sensor presented a bulk RI sensitivity of 1943 nm/RIU, while the as-prepared sensor exhibited a value of 4561 nm/RIU. The immobilization of the Cu3-xP@Cu-BTC interface enhanced the bulk RI sensitivity of the sensor by 2.35 times. It indicated that the interaction strength between the optical microfiber and surrounding matter increased by 2.35 times at the locations where the evanescent field reached. The taper structure of microfiber excited the HE12 mode, except the fundamental HE11 mode in the sensing region [20]. The two modes interfered and created an interferometric fringe in the fiber transmission spectrum. A numerical mode simulation (COMSOL) software was utilized to analyze the transverse electric field amplitude distribution of the HE12 mode, which distributed into the surface coating, sensed the change in the surface RI, and translated it into wavelength shifts (Fig. 4B). For an optical microfiber with waist diameter of 6.4 µm, the penetration depth was about 800 nm. The development of interface with high RI pulled more energy from the core of the fiber to the surface, thus, forming the evanescent field, which enhanced the interaction between microfiber and the surrounding (Fig. 4B and Fig. S8; the thickness and optical constants such as RI and extinction coefficient of Cu-BTC, Cu3-xP and Cu3-xP@Cu-BTC were determined by spectroscopic ellipsometry.). That is, the RI sensitivity of microfiber increased by 2.35 times in this 800 nm space. However, as a sensor monitoring the surface RI changes induced by the target binding, the improvement of surface RI sensitivity is more important than that of the bulk RI sensitivity. Thus, it requires not only the enhancement of energy, but also the confinement of it on the microfiber surface.

Fig. 4

A) The RI sensitivity of the silica microfiber and microfiber with Cu3-xP@Cu-BTC interface, B) The transverse electric field amplitude distributions of HE12 mode of the silica microfiber and microfiber with Cu3-xP@Cu-BTC interface. The calculated near-field intensity map of the microfiber surfaces C) without interface, D) with Cu3-xP interface and E) with Cu3-xP@Cu-BTC interface. F) The localized electric field enhancement by different interfaces.

Fig. 4C-4E and S9 present the calculated near-field intensity of the microfiber surface with different interfaces through a finite-difference time-domain (FDTD) approach. During the calculation, the 3D Cu3-xP@Cu-BTC interface was artificially split into several 2D layers consisting of 3 nm of Cu-BTC and a single-layer of Cu3-xP for modeling convenience. The strongest electric field enhancement (|E/E|) on the surfaces with different interfaces are graphed in Fig. 4F. The plasmonic Cu3-xP interface enhanced the evanescent field on microfiber surface by 14 times in two dimensions. The incorporation of the Cu-BTC framework led to the plasmonic resonance enhancement of Cu3-xP in three dimensions, with an increment of 20 times every single layer. For the Cu3-xP@Cu-BTC interface, the 3D coupling between the 2D layers significantly enhanced the electric field, which resulted in the overall electric field enhancement of more than 20 times. Moreover, the plasmonic resonance enhancement extended to three dimensions increased the contact possibility of nanoparticles/molecules with the sensor surface with strong electric field. Thus, the electric field enhancement effect of the 3D Cu3-xP@Cu-BTC interface is significant than the 2D Cu3-xP interface. Assisted by the LSPR of 3D nanointerface, the exterior mode showed a high localization, and was concentrated in the exterior surface of microfiber.

Acting as a wavelength-encoded sensor, the optical microfiber sensor recorded the spectral shift induced by the binding of the nanoparticles on its surface. Treating as a point-like distortion, the change in the propagation constant of each mode induced by the single-particle binding event at the microfiber pan class="Gene">surface could be expressed as [38] where α is the polarizability of the particle in excess of the medium, E(x, y) is the localized electric field at the binding site, V is the volume of a single particle, ε is the permittivity, E(l) is the electric field of microfiber, Sur is the microfiber surface area, and L is the length of microfiber waist. Thus, the wavelength shift of the sensor induced by the single-particle binding event was proportional to the surface energy at the binding site. When the nanoparticle binding occurred at the Cu3-xP@Cu-BTC interface, the induced wavelength shift would be more than 400-fold larger than the system without interface. Therefore, compared with sensors monitoring the bulk RI changes, the 3D nanointerface-functionalized microfiber is more suitable for monitoring the binding events on its surface, due to the evanescent field confinement and surface enhancement.

Single-nanoparticle detection

The single-particle detection ability of the as-prepared sensor is shown in Fig. 5 compared with two control microfiber sensors, including a microfiber sensor without interface (Fig. 5A) and a sensor with Cu3-xP interface (Fig. 5B). It was observed that the sensors appeared stable in the blank solution (pan class="Chemical">ethanol) (Fig. 5A-5C). As the spherical polystyrene (PS) nanoparticles solution was injected in ethanol to form a suspension with a concentration of 2.5 × 103 particles/mL, the silica microfiber without interface exhibited an optical response similar to the blank solution (Fig. 5A), while the microfiber sensor with the Cu3-xP interface recorded a general redshift in the transmission spectrum in the first 8 min (Fig. 5B). Thus, the sensor without interface did not recognize the single-nanoparticle-induced signal. According to the resolution of OSA (0.02 nm), the step change of 0.02 nm in Fig. 5B too did not recognize the single-nanoparticle-induced signal. In the case of the as-prepared sensor, a stepwise shift of about 0.2 nm was recorded 5 min after the injection of the nanoparticle dispersion (Fig. 5C). It indicates that for the nanoparticles with a diameter of 170 nm, the single-particle binding signal could be recognized by the as-prepared sensor. Fig. 5C to 5E show that the probability of the single-particle binding events increased with solution concentration. Also, the total wavelength shift in the spectrum increased as the concentration of the injected dispersion increased (Fig. 5F). The CCD images in Fig. 5G to 5I confirmed that the nanoparticles attached to the sensor surface one by one, instead of clusters. Thus, the as-prepared sensor could recognize the single-nanoparticle-binding signal due to the evanescent field enhancement of the plasmonic Cu3-xP@Cu-BTC interface, while the two control sensors were not able to detect such signal.

Fig. 5

Single-nanoparticle detection. Optical response of the A) control sensor without interface, B) control sensor with only Cu3-xP interface, and C) as-prepared sensor (concentration of solution: 2.5 × 103 particles/mL). Optical response of the as-prepared sensor to particles at a concentration of D) 2.5 × 104 particles/mL and E) 2.5 × 105 particles/mL. F) The total wavelength shift in the spectrum with the increasing-concentration solutions injecting. G-I) CCD images of the nanoparticles of D = 170 nm deposited on the microfiber surface. (concentration in case of G: 2.5 × 103 particles/mL, H: 2.5 × 104 particles/mL and I: 2.5 × 105 particles/mL).

Single-nanoparticle detection. Optical response of the A) control sensor without interface, B) control sensor with only Cu3-xP interface, and C) as-prepared sensor (concentration of solution: 2.5 × 103 particles/pan class="Gene">mL). Optical response of the as-prepared sensor to particles at a concentration of D) 2.5 × 104 particles/mL and E) 2.5 × 105 particles/mL. F) The total wavelength shift in the spectrum with the increasing-concentration solutions injecting. G-I) CCD images of the nanoparticles of D = 170 nm deposited on the microfiber surface. (concentration in case of G: 2.5 × 103 particles/mL, H: 2.5 × 104 particles/mL and I: 2.5 × 105 particles/mL).

Optical sizing

In order to explore the ability of the as-prepared sensor to distinguish the nanoparticles with different diameters, four types of nanoparticles with diameters of 50 nm, 80 nm, 100 nm and 500 nm were employed. As shown in Fig. 6 , as the solutions of nanoparticles with different diameters were respectively injected into the blank solution to form a solution with a concentration of 2.5 × 103 particles/mL, similar stepwise transmission shifts were recorded by the as-prepared sensor. For instance, for nanoparticles with diameter of 50 nm, the as-prepared sensor presented a stepwise shift of 0.05 nm in the transmission spectrum (Fig. 6A). For nanoparticles with a diameter of 80 nm, it exhibited a stepwise shift of 0.075 nm (Fig. 6B). For nanoparticles with a diameter of 100 nm, it exhibited a stepwise shift of 0.10 nm (Fig. 6C). On the other hand, for nanoparticles with a diameter of 500 nm, it showed a stepwise shift of 0.60 nm (Fig. 6D). It was noted that the single-nanoparticle-induced step change in the transmission spectrum increased with the nanoparticle size. The stepwise wavelength shifts corresponding to the single-nanoparticle binding associated with different diameters are demonstrated in Fig. 6E, which indicated a sensitivity of 0.00125 nm (stepwise shift)/nm (diameter). The limit of detection (LOD) (3σ/s) [39] was 48 nm. Similar results have been reported for other optical sizing methods [40]. Further, when the sensor was exposed to the mixture with the above particles at a concentration of 2.5 × 103 particles/pan class="Gene">mL, it recorded a stepwise shift of 0.10 nm in the first 150 s and a stepwise shift of 0.60 nm in 1500 s, respectively, indicating that individual particle with diameters of 100 nm and 500 nm bound to the fiber (Fig. 6F). The SEM images of the sensor surface four weeks after functionalization in Fig. S10 also reveal the reusability of the sensor.

Fig. 6

Optical response of the as-prepared sensor to various particle sizes at a concentration of 2.5 × 103 particles/mL: A) D = 50 nm, B) D = 80 nm, C) D = 100 nm, and D) D = 500 nm. E) Stepwise wavelength shifts corresponding to the single-nanoparticle binding associated with different diameters. F) Optical response of the mixed solution with various particle sizes (sizes of particles: D = 80 nm, 100 nm, 170 nm and 500 nm).

Potential in detecting ultrafine particulate pollutants

In order to investigate the practical potential in detecting ultrafine particulate pollutants, the as-prepared sensor was employed to detect the ultrafine carbon powders. The stepwise optical response of 0.2 nm in Fig. 7 A indicates that there were a particle with diameter of about 200 nm coming to the sensor surface. The morphology ultrafine carbon powders in SEM image in Fig. 7B confirms the sizing result obtained in Fig. 7A. This shows the practical potential of sensor in detecting ultrafine particulate pollutants.

Fig. 7

A) Optical response of the as-prepared sensor to the ultrafine carbon powder dispersion solution (concentration: 10-12 mg/mL). B) SEM image of the ultrafine carbon powders.

Conclusions

In conclusion, a portable silica optical microfiber sensor with low-energy supply for single nanoparticle detection and sizing has been developed, by real-time monitoring the changes in pan class="Gene">surface RI-induced transmitted spectrum. The sensor was functionalized by a 3D plasmonic nanointerface made of Cu-BTC framework supporting the Cu3-xP nanocrystals. The LSPR of the Cu3-xP nanocrystals was tuned to coincide with the evanescent field of the optical microfiber. The 3D Cu-BTC framework, as the scaffold of nanocrystals, assisted the confinement of the local resonance field on the microfiber with three dimensions, at which the binding of the target nanoparticles occurred. Based on the energy confinement and surface enhancement of the Cu3-xP@Cu-BTC nanointerface, the optical microfiber sensor overcame its sensitivity limit and enabled the detection and sizing of the individual nanoparticles. The compact size with low power supply of the sensor underline its suitability as a portable device for the real-time single-nanoparticle detection, especially for the convenient evaluation of the ultrafine particles in the environment. This work opens up new pathways for the portable real-time single-nanoparticle characterization in the future, along with shedding light on the effective and universal approach to overcome the sensitivity limit of the optical microfibers through the energy control by an interfacial design.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.