Introduction

Carbon nanotubes (CNTs) have become very interesting systems due to their unique structural and electronic properties as well as their diverse applications [1]. Since the discovery by Iijima and Ichihashi, CNTs have drawn great attention due to their superior properties such as high surface area and chemical stability [2]. The surface of CNTs congregates more negative charges and has been actively and widely studied [3]. Historically, CNTs have acted as probe molecule to be used in sensitive Raman test and study about the interfacial reactions between CNTs and metallic support for their unique electronic features, high aspect ratio and strong electron–phonon coupling [4]. Doping of carbon nanotubes was first attempted by Stephan et al. in 1994; they synthesized boron-doped CNTs using arc-discharge method [5]. Raman spectroscopy shows that boron doping of CNTs converted their chemical and optical properties [6]. It has been shown that boron doping in nanotubes gives rise to a prominent acceptor strip near the Fermi level, and as a result the nanotubes are plentiful in holes. Therefore, the presence of the acceptor band ensures that the boron-doped CNTs display permanent p-type character [7].

Detecting single molecules and identifying their chemical structures are among the main scopes of increased interest. The observation of spectroscopic signals at the single molecule level is the main scientific advantage, as it provides insight into the intrinsic properties of a molecule [8]. Fluorescence microscopy is used as a touchy and single molecule spectroscopic tool which deals detecting single molecules by fluorescence and it must be demonstrated by fluorescent dye molecules to get high enough fluorescence quantum yields and detectable spectral properties [9]. However, in SERS technique, fluorescence is a disturbing process and with some experimental techniques, this problem should be resolved. In SERS when the target molecule is joined to noble metallic nanoparticles (e.g., Ag, Cu, Au,…), the Raman signal can be enhanced up to more than tenfold of magnitude [10, 11]. Silver nanoparticles are the only metal whose plasmon resonance can be adapted to any wavelength in the visible spectrum [12, 13]. Flemish in 1974 used harsh Ag electrode surface to record the first SERS spectrum [14]. The technique of SERS can be applied for obtaining the materials structural and identification information in low concentrations [15]. In this article, for confirming the high SERS activity of Ag NPs, the B-CNTs composite has been chosen as a probe because it has been well characterized by SERS and most of the prominent Raman bands have been assigned. For explaining the enhanced mechanism of SERS, it is now generally accepted that there are two interacting clarifications, which explain the origin of SERS effect: electromagnetic [16] and chemical effects [17]. The cross sections of spontaneous Raman scattering (~ 10−30 cm−2) are much smaller than that of IR absorption (~ 10−19 cm−2) or fluorescence (~ 10−16 cm−2). Therefore, signal enhancement technique is very useful for Raman scattering measurements. Raman scattering from molecules adsorbed on metallic nanostructures is strongly enhanced due to the excitation of localized surface plasmon polaritons. This phenomenon is known as surface-enhanced Raman scattering (SERS), the enhancement factor of which has been reported to reach up to 105–1020 [18]. The organization of this article is as follows: (1) a new method for introducing boron oxide (B2O3) molecule to the carbon nanotubes lattice structure with laser ablation in a liquid environment (LALE) will be explained. (2) Silver nanoparticles (Ag NPs) will be fabricated via Lee–Meisel method. (3) Spontaneous and SERS spectra of pure and B-doped carbon nanotubes will be presented by experimental and simulation methods. (4) Finally, the results of the experimentally recorded spontaneous and SERS Raman spectra will be compared by the relevant spectra obtained by the simulation method.

Materials and methods

Preparation and SEM imaging of CNT/B2O3

The carbon nanotubes (CNTs) (diameter ~ 15–20 nm; length ~ 2.5 μm) were prepared by CVD method grown on Ni catalyst at 660 °C for 20 min. The CNT/B2O3 was prepared by using laser ablation method using the fundamental wavelength of a Ce-doped Nd/YAG laser at 1064 nm. First, pure CNTs powder was mixed with boron oxide (B2O3), with a ratio of 97:3 wt%, respectively. Then, the sample was treated in ethanol solution and the suspending liquid using laser beam (80 mJ) for 15 min. The mixture solution was heated and stirred during laser irradiation. The sample was dispersed under ultra-sonication technique for 10 min to produce uniformed CNT/B2O3. The resulting material is shown in Fig. 1 in a representative SEM image of doped CNTs (CNT/B2O3). The scale bar is 200 nm.

Fig. 1
figure 1

SEM image of doped CNTs (CNT/B2O3)

Full size image

Preparation of Ag NPs

Ag NPs because of their ability in exhibiting localized surface Plasmon resonances (LSPR) have found many applications in bio-detection LSPR excitation can considerably enhance the local electric field which is the main reported mechanism of surface-enhanced Raman spectroscopy (SERS) [19]. A variety of methods for the synthesis of Ag NPs has been reported in the scientific literature. This preparation is simple, but great care must be exercised to make stable and reproducible colloids. The wet chemical syntheses can produce nano-sized colloids in many different shapes [20]. The processes can be divided into two categories (traditional and non-traditional). The traditional methods are the solution-phase synthesis techniques that are based on different modifications of the Lee–Meisel method, and there are a series of examples in which different Ag salts and reducing factors are used [12, 13]. The non-traditional methods include vapor-phase condensation of a metal onto a solid support [21,22,23], laser ablation of a metal target into a suspending liquid [24], photoreduction of Ag ions [25, 26] and electrolysis of an Ag salt solution [27, 28]. With the solution-phase synthesis techniques, the main problem is a limited flexibility in the size of particles that can be produced. For Ag NPs suspensions, a common method is the Lee–Meisel method, which is a variation of the Turkevich method in which AgNO3 is used as the metal source [29].

In this work, silver nanoparticles were prepared using the Lee–Meisel method. AgNO3 (17 mg) was dissolved in 100 mL of deionized water and brought to boiling. A solution of 1 wt% sodium citrate (7 mL) was added to the reaction solution. The solution was kept on boiling for 8 min and cooled to room temperature.

The chemical reaction of the process is as follows:

$$4{\text{Ag}}^{ + } + {\text{ C}}_{6} {\text{H}}_{5} {\text{O}}_{7} {\text{Na}}_{3} + 2{\text{H}}_{2} {\text{O }}\rightarrow 4{\text{Ag }} + {\text{ C}}_{6} {\text{H}}_{5} {\text{O}}_{7} {\text{H}}_{3} + \, 3{\text{Na}}^{ + } + {\text{O}}_{2} .$$

Results and discussion

Experimental section

Absorption spectroscopy and SEM imaging of Ag NPs

Ultraviolet–visible-near infrared (UV/Vis/NIR) spectroscopy refers to absorption or reflectance spectroscopy in this region of electromagnetic waves. This means it uses light in the visible and adjacent ranges. The absorption or reflectance in the visible range has direct impacts on the perceived color of the chemicals involved. Absorption spectroscopy is complementary to fluorescence spectroscopy, in that fluorescence deals with transitions from the excited state to the ground state, while absorption gauges transitions from the ground state to the excited state [30]. UV/Vis/NIR absorption spectroscopy was used to provide further insight into the interaction among ground state electrons within the hybrid structures [31].

The localized surface plasmon resonance (LSPR) of Ag nanoparticles is illustrated in the UV/Vis/NIR absorption spectra collected by UV/Vis/NIR spectrometer manufactured by T80+ PG Instruments, and a typical spectrum is shown in Fig. 2. There is an absorption peak at ~ 420 nm for Ag NPs.

Fig. 2
figure 2

UV/Vis/NIR absorption spectrum of the Ag NPs synthesized by the Lee–Meisel method

Full size image

By increasing the size of the silver nanoparticles, the absorption spectrum shifts to higher wavelengths. The observed spectral shift results from the distributing of the particles surface charge upon a larger surface area so that the surrounding medium make up better the restoring force thus slowing the electron oscillations [32]. According to the absorption spectrum obtained from silver nanoparticles (420 nm), the synthesized nanoparticles have an average grain size of around 50 nm [33].

SEM image of Ag NPs is shown in Fig. 3, and the average sizes of the Ag NPs are around 50 nm. The scale bar is 200 nm.

Fig. 3
figure 3

SEM image of the Ag NPs synthesized by the Lee–Meisel method

Full size image

Raman spectroscopy of CNT

Raman spectroscopy was used to investigate the crystallinity and the structural changes of carbon framework of CNTs. To perform Raman spectroscopy, Almega Dispersive Raman spectrometer system manufactured by Thermo Nicolet with the second harmonic of a Nd/YLF laser beam with a wavelength of 532 nm was used. As shown in Fig. 4, the CNTs sample shows prominent peaks at 1346 cm−1 (D band) and also a peak at 1568 cm−1 (G band). The intensity of the vibrating D mode (which shows the presence of deformation, impurities and amorphous carbons in the carbon nanotubes) and the intensity of the vibrating G mode (which displays the existence of the order in the carbon network) were examined. The D band represents edges, other defects and disordered carbon due to the vibration of sp3-bonded carbon atoms and impurities, whereas the G band arises from the zone center E2g mode, assigning to the ordered sp2-bonded C atoms [34].

Fig. 4
figure 4

Raman spectrum of CNT synthesized by CVD method

Full size image

The Raman intensity of D and G peaks of CNT powder is 1800 (a.u.) and 2400 (a.u.), respectively. The degree of crystallinity in carbon samples was evaluated using the D–G band intensity (ID/IG) ratio, which is the intensity of the disorder mode at D band divided by the intensity of graphite mode at G band. As shown in Fig. 4, due to the purity of the sample, the intensity of the Raman spectrum of (ID/IG) ratio is less than one.

$$\left( {I_{\text{D}} } \right)_{\text{N}} = 1800\;\left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{G}} } \right)_{\text{N}} = 2400\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\frac{{I_{\text{D}} }}{{I_{\text{G}} }} = \frac{1800}{2400} = 0.75$$

Raman spectroscopy of CNT/B2O3

The predominant band in the spectrum of B2O3 is located at 850 cm−1 [35,36,37]. As shown in Fig. 5, the Raman intensity of D and G peaks of CNTs/B2O3 is at around 700 (a.u.) and 450 (a.u.), respectively. The ID/IG ratio after introducing B2O3 is higher than that of pure CNTs, which can be related to the doping effect

$$\left( {I_{\text{D}} } \right)_{\text{N}} = 700 \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{G}} } \right)_{\text{N}} = 450 \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{{{\text{B}}_{2} {\text{O}}_{3} }} } \right)_{\text{N}} = 125 \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\frac{{I_{\text{D}} }}{{I_{\text{G}} }} = \frac{700}{450} = 1.5$$

Increasing ID/IG value from 0.75 to 1.5 suggests a decrease in in-plane size of graphene upon the reduction process [38]. It can be seen that the Raman intensity of vibrational mode at 850 cm−1 (corresponding to the B2O3) is 125 (a.u.).

Fig. 5
figure 5

Raman spectrum of CNT/B2O3

Full size image

Raman enhancement of CNT/B2O3–Ag NPs

In this section, we have compared the Raman intensities of CNT/B2O3, CNT/B2O3–Ag NPs. As shown in Fig. 6, we can see that the Raman intensity of CNT/B2O3–Ag NPs is larger than that of CNT/B2O3. For Raman peak position at 850 cm−1 (B2O3), the Raman intensity of B2O3 is about 125 (a.u.) and for B2O3–Ag NPs is about 2400 (a.u.). The reason can be considered due to the intensity enhancement by Ag NPs.

Fig. 6
figure 6

Raman spectrum of CNT/B2O3–Ag NPs

Full size image
$$\left( {I_{\text{D}} } \right)_{\text{SERS}} = 9000\;\left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{G}} } \right)_{\text{SERS}} = 3500\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{{{\text{B}}_{2} {\text{O}}_{3} }} } \right)_{\text{SERS}} = 2400\; \left( {{\text{a}} . {\text{u}} .} \right)$$

The SERS enhancement factor (EF) has been defined [39]:\({\text{EF}} = \frac{{I_{\text{SERS}} }}{{I_{\text{Normal}} }} \times \frac{{V_{\text{Normal}} }}{{V_{\text{SERS}} }} \times \frac{{C_{\text{Normal}} }}{{C_{\text{SERS}} }}\), where ISERS and INormal are the peak intensity on CNT/B2O3–Ag NPs and CNT/B2O3, respectively; VSERS and VNormal are the probe volumes for the CNT/B2O3–Ag NPs and CNT/B2O3 measurements, respectively; and CSERS and CNormal are the CNT/B2O3 concentrations that contribute to SERS (ISERS) and normal Raman intensities (INormal), respectively (Fig. 7).

Fig. 7
figure 7

A comparison between SERS, left, and spontaneous Raman, right, spectra presented in Figs. 5 and 6

Full size image

For D, G and B2O3 peaks, the EF is:

$$\left( {\text{EF}} \right)_{\text{D}} = \frac{{I_{\text{SERS}} }}{{I_{\text{Normal}} }} \times \frac{{V_{\text{Normal}} }}{{V_{\text{SERS}} }} \times \frac{{C_{\text{Normal}} }}{{C_{\text{SERS}} }} = \frac{9000}{700} \times 1 \times 1 = 12.8$$
$$\left( {\text{EF}} \right)_{\text{G}} = \frac{{I_{\text{SERS}} }}{{I_{\text{Normal}} }} \times \frac{{V_{\text{Normal}} }}{{V_{\text{SERS}} }} \times \frac{{C_{\text{Normal}} }}{{C_{\text{SERS}} }} = \frac{3500}{400} \times 1 \times 1 = 8.8$$
$$\left( {\text{EF}} \right)_{{{\text{B}}_{2} {\text{O}}_{3} }} = \frac{{I_{\text{SERS}} }}{{I_{\text{Normal}} }} \times \frac{{V_{\text{Normal}} }}{{V_{\text{SERS}} }} \times \frac{{C_{\text{Normal}} }}{{C_{\text{SERS}} }} = \frac{2400}{125} \times 1 \times 1 = 19.2$$

The synthesis of Ag NPs can be used as a stable suspension for the production of enhanced Raman signals for up to 1 month. The intensity of the SERS signals is enhanced by up to 12-fold. The enhancement can be related to the concentration of Ag NPs and the molarity of two probe molecules [40]. The IB2O3, ID and IG of CNT, CNT/B2O3 and CNT/B2O3–Ag NPs samples are summarized in Table 1.

Table 1 ID, IG and \(I_{{{\text{B}}_{2} {\text{O}}_{3} }}\) of CNT, CNT/B2O3 and CNT/B2O3–Ag NPs samples from the spectra presented in Figs. 2, 3 and 4
Full size table

Simulation section

In the simulation approach, using Hyperchem, Gauss View and Gaussian software was drawn pure CNT, boron-doped CNT and CNT/boron–Ag NPs. Then, the Raman spectra were obtained using the DFT (density functional theory) and B3LYP (Becke–Style three-parameter) computing methods and the 6-31G base set.

Raman spectroscopy of CNT

The schematic illustration and Raman spectrum of pure carbon nanotube are presented in Fig. 8a, b, respectively.

Fig. 8
figure 8

a Schematic illustration of CNT structure. b Theoretical calculation of Raman spectrum of CNT

Full size image

The Raman intensity of D and G peaks of CNT is 18 (a.u.) and 80 (a.u.), respectively. As can be seen, due to the purity of the sample, the intensity of the Raman spectrum of (ID/IG) ratio is less than one.

$$\left( {I_{\text{D}} } \right)_{\text{N}} = 18\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{G}} } \right)_{\text{N}} = 80\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\frac{{I_{\text{D}} }}{{I_{\text{G}} }} = \frac{18}{80} = 0.22$$

Raman spectroscopy of CNT/boron

The schematic illustration and the Raman spectrum of boron-doped carbon nanotube are shown in Fig. 9a, b, respectively.

Fig. 9
figure 9

a Schematic illustration of boron-doped CNT. b Theoretical calculation of Raman spectrum of boron-doped CNT

Full size image
$$\left( {I_{\text{D}} } \right)_{\text{N}} = 38\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{G}} } \right)_{\text{N}} = 78\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{Boron}} } \right)_{\text{N}} = 8\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\frac{{I_{\text{D}} }}{{I_{\text{G}} }} = \frac{38}{78} = 0.48$$

The Raman intensity of D, G and boron peaks CNT are 38 (a.u.), 78 (a.u.) and 8 (a.u.), respectively. As can be seen, due to boron-doped CNT, the intensity of the Raman spectrum of (ID/IG) ratio has increased.

Raman enhancement of CNT/boron–Ag NPs

The schematic illustration and Raman spectrum of CNT/boron–Ag NPs are shown in Fig. 10a, b, respectively.

Fig. 10
figure 10

a Schematic illustration of CNT/boron–Ag NPs. b Theoretical calculation of Raman spectrum of CNT/boron–Ag NPs

Full size image
$$\left( {I_{\text{D}} } \right)_{\text{SERS}} = 510\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{G}} } \right)_{\text{SERS}} = 250\; \left( {{\text{a}} . {\text{u}} .} \right)$$
$$\left( {I_{\text{Boron}} } \right)_{\text{SERS}} = 520\; \left( {{\text{a}} . {\text{u}} .} \right)$$

The Raman intensity of D, G and boron peaks CNT are 510 (a.u.), 250 (a.u.) and 520 (a.u.), respectively. Ag NPs created a multitude of hot spots, with a strong SERS effect. The Raman scattering of CNT/boron has been enhanced as a result of a strongly enhanced electric field due to the presence of Ag NPs [41].

$$\left( {\text{EF}} \right)_{\text{D}} = \frac{{I_{\text{SERS}} }}{{I_{\text{Normal}} }} \times \frac{{V_{\text{Normal}} }}{{V_{\text{SERS}} }} \times \frac{{C_{\text{Normal}} }}{{C_{\text{SERS}} }} = \frac{510}{38} \times 1 \times 1 = 13$$
$$\left( {\text{EF}} \right)_{\text{G}} = \frac{{I_{\text{SERS}} }}{{I_{\text{Normal}} }} \times \frac{{V_{\text{Normal}} }}{{V_{\text{SERS}} }} \times \frac{{C_{\text{Normal}} }}{{C_{\text{SERS}} }} = \frac{250}{78} \times 1 \times 1 = 3.2$$
$$\left( {\text{EF}} \right)_{\text{Boron}} = \frac{{I_{\text{SERS}} }}{{I_{\text{Normal}} }} \times \frac{{V_{\text{Normal}} }}{{V_{\text{SERS}} }} \times \frac{{C_{\text{Normal}} }}{{C_{\text{SERS}} }} = \frac{520}{8} \times 1 \times 1 = 65$$

The simulation results of the intensities for Iboron, ID and IG of CNT, CNT/boron and CNT/boron–Ag NPs samples are summarized in Table 2.

Table 2 The simulated intensities for Iboron, ID and IG of CNT, CNT/boron and CNT/boron–Ag NPs samples from the spectra in Figs. 8, 9 and 10
Full size table

Table 3 shows the comparison between the results of the experimental and the simulation methods.

Table 3 The comparison of the enhancement factor (EF) for boron oxide and also for the D and G modes results for the experimental and the simulation approaches
Full size table

Conclusions

In this article, a facile procedure to synthesize CNT/B2O3 hybrid and its spontaneous Raman and the signal enhancement at Ag nanostructured surface using SERS has been presented.

The increase in the Raman signal intensity (for experimental and simulation approaches) reveals the usefulness and effectiveness of the SERS method for detecting specific molecules with low concentrations in a matrix environment on the surface of Ag NPs. In order to compare the obtained EFs, it is also possible to use other noble metallic nanoparticles such as gold, copper and titanium to enhance the Raman signal.