Introduction

Because of its outstanding physicochemical, ferroelectric, piezoelectric, and nonlinear optical capabilities, lithium niobate is one of the interesting and promising ferroelectric materials [1,2,3]. Due to these properties, it has been widely applied in a variety of fields, including optical waveguides, electro-optical modulation, holographic storage, optical parametric oscillators, etc. [4, 5]. It belongs to the space group R3c and has rhombohedral symmetry. LN’s ferroelectric phase changed to a paraelectric phase at 1200 °C. Due to its exceptional qualities, LiNbO3 has been widely used as a promising contender for “optical silicon.” Nevertheless, LN serves as a passive component and an insulator in the aforementioned applications [6,7,8]. Many of the remarkable features of LN can be utilized in semiconductor integrated devices if LN devices can be integrated onto a silicon substrate to produce an active component, combining the advantages of both LN and semiconductor materials [9,10,11]. The main challenges in fabricating an active component based on the p–n junction were the low electrical conductivity and optical absorption of LN. In an attempt to fabricate a device using LN, it is necessary to increase the electrical conductivity of LN [12,13,14]. As reported, various methods were employed to synthesize LiNbO3, for example, solution phase, sol–gel, citrate gel, electrospinning, and pulsed laser deposition [15,16,17,18]. There are many routes used to control the optical and electrical properties of LN, such as doping with different elements, for example Mg, Zn, Fe, etc. [19,20,21,22]. Several attempts to synthesize lithium niobate with a gold and silver core/shell structure have been reported. Murillo et al. [2] deposited Ag–LiNbO3 nanocomposite thin films with Ag nanoparticles (NPs) embedded into the LN matrix by the co-deposition of Ag and LN using a pulsed laser deposition (PLD) method that shows a surface plasmon resonance (SPR) in the visible range. The fabrication and characterization of a Fe-doped LiNbO3/n-Si heterojunction was demonstrated by Li et al. [23]. As reported, on solid surfaces, gold nanoparticles were assembled into 2D or 3D superlattices, nanowires, colloidal aggregates, and a core/shell in order to improve their physical and chemical properties and enable them to be efficient for many important applications such as catalysts, sensors, photonics, and surface-enhanced Raman scattering [24,25,26,27,28,29,30]. Compared to other techniques, the laser ablation of Au@LiNbO3 core/shell nanoparticles in liquid method used is promising and has several advantages including simplicity, low-cost, high purity product, fast, green synthesis, and fair control of the core–shell dimensions. [31,32,33]. The core–shell morphology give better thermal stability and decreases the reactivity.

Herein, we have demonstrated the synthesis of Au@LiNbO3 core/shell nanoparticles by two- step laser ablation in methanol at different laser fluences. The fabrication of Au@LiNbO3/Si was proposed as a simple, novel, and high-performance photodetector can used for many technological applications.

Experimental Work

Preparation of Au@LiNbO3 Core–Shell Nanoparticles

Colloidal Au nanoparticles are prepared by irradiating a high purity (99.9%) Au pellet immersed in a vessel filled with 3 ml of ethanol with 70 laser pulses of a second harmonic generation (= 532 nm) Q-switched Nd:YAG laser with a pulse duration of 7 ns and a repetition frequency of 1 Hz to synthesize Au@LiNbO3 core/shell nanoparticles. The ablation of Au nanoparticles was carried out at a laser fluence of 1.5 J/cm2. After Au colloidal preparation, a LiNbO3 wafer with an area of 1 cm2 was placed in the quartz vessel filled with Au colloid and then irradiated with 300 laser pulses at a laser fluence of 1.3, 1.6, 2, and 2.2 J/cm2. The experimental setup of the PLAL system is shown in Fig. 1a.

Fig. 1
figure 1

a Schematic diagram of detailed formation mechanism of Au @ LiNbO3 core/shell nanoparticles by pulsed laser ablation in ethanol b Fabrication steps of Au@LN/Si photodetector

Full size image

Characterization of Au @ LiNbO3 Core/Shell Nanoparticles

The optical absorption of colloidal Au nanoparticles, LN nanoparticles, and Au@LiNbO3 core/shell nanoparticles were measured using a UV–Vis spectrophotometer (Shimadzu UV-2550). The structural properties of the Au@LiNbO3 core/shell were investigated using an X-ray diffractometer (Panalytical X’Pert Pro). A confocal Raman spectrometer was used to record Raman spectra (HORIBA XPLORA PLUS). Transmission electron microscopy (ZEISS LEO 912) was used to examine the morphology and size of the Au@LiNbO3 core–shell.

Fabrication and Characterization of Au@LiNbO3/Si Photodetector

Fabrication of Au@LiNbO3/p-Si heterojunction photodetector was performed by depositing a thin-film Au@LiNbO3 core/shell via a spin coating route on polished (111)-oriented p-type silicon substrate. The silicon substrates used here have areas of 1 cm2 and electrical resistivity of 1–3 Ωcm. The ohmic contacts were made by the depositing of Al and In electrodes on the Au@LiNbO3 nanostructure film and silicon substrate, respectively, by using the thermal resistive technique through a metal mask. Figure 1b shows the fabrication steps of the Au@LiNbO3/Si photodetector. The current–voltage characteristics of the heterojunctions in the dark and under illumination were measured. The capacitance–voltage characteristics of the heterojunction were measured at a frequency of 200 kHz using LCR meter. The responsivity of the photodetector was measured using a monochromator (Jobin Yovn) in the spectral range of 370–1000 nm at a bias voltage of − 8 V. The power calibration of the monochromator was performed using a silicon power meter.

Results and Discussion

Figure 2 shows the XRD pattern of Au@LN core/shell nanoparticles prepared at 1.3 J/cm2, which exhibits two peaks located at 2θ = 33.05° and 61.6° which correspond to the (104) and (214) planes, respectively. The XRD pattern of the nanoparticles prepared at 1.6 J/cm2 shows the presence of eight XRD peaks located at 2θ = 33.03°, 35.9°, 41.08, 47.7°, 54.6°, 56.3°, 57.3°, and 61.7°, which correspond to (104), (110), (113), (024), (116), (122), (214), and (220) planes, respectively. While the XRD patterns of the core/shell synthesized at 2 J/cm2 have six diffraction angles at 2 = 33.04°, 35.84°, 47.7°, 54.5°, 56.6°, 57.3°, and 61.7°, corresponding to planes (104), (110), (024), (116), (122), and (214), respectively. Finally, the nanoparticles synthesized at 2.2 J/cm2 show the presence of six diffraction peaks observed at 2θ = 33.06°, 35.9°, 47.8°, 54.6°, 56.4°, 57.3°, and 61.7° corresponding to (104), (110), (024), (116), (122), and (214) planes, respectively. All the observed XRD peaks are indexed to a rhombohedral LiNbO3 according to JCPDs#20- 0631 [34,35,36]. The XRD peaks of core Au nanoparticles were observed at 2θ = 44° and 66° which correspond to (200) and (220) planes, respectively.

Fig. 2
figure 2

XRD patterns of Au@LN core/shell nanoparticles synthesized different laser fluences

Full size image

Figure 3 depicts the UV–Vis absorption spectra of colloidal Au@LN core and shell prepared at various laser fluences. The optical absorption spectra display a general increase with increased laser fluence, besides two groups of absorption peaks, the first one is attributed to LiNbO3 that is located at 224, 226, 227, and 231 nm for nanoparticles synthesized at 1.3, 1.6, 2, and 2.2 J/cm2 laser fluence, respectively, with the shoulder diminishing at higher laser fluence. The second peak is located at ~ 521–530 nm, which belongs to the surface plasmon resonance (SPR) of the gold nanoparticles. This peak disappeared at laser fluence of 2.2 J/cm2 This may be due to an increase in shell thickness. Increasing the LN shell thickness leads to decreasing the intensity of SPR of the Au core [37, 38]. The increase in LN concentration at higher laser fluence may be attributed to the low dipole moment of ethanol (1.69D) supporting growth via introducing extra species during the ablation, which results in a wider particle size distribution [39, 40]. Figure 4 displays the optical absorption of Au nanoparticles and the change in optical absorbance of LN nanoparticles after making the core/shell structure with Au nanoparticles.

Fig. 3
figure 3

UV–Vis absorbance of Au/LiNbO3 ablated in ethanol at different laser fluence

Full size image
Fig. 4
figure 4

UV–Vis absorbance of Au@LiNbO3, colloidal Au nanoparticles, and LiNbO3 nanoparticles

Full size image

The optical energy gap of LiNbO3 and Au@ LiNbO3 core/shell nanoparticles was estimated using Tauc’s relationship [41,42,43]. The optical energy gap value is estimated by plotting (αhv)2 as a function of hν, and the extrapolation of the second region of the curve to the photon energy axis gives the energy gap, as shown in Fig. 5. The energy gaps of the Au@LN core–shell prepared at laser fluence of 1.3, 1.6, 2, and 2.2 J/cm2 were 3.6, 3.49, 3.4, and 3.8 eV, respectively. The energy gap of LiNbO3 was 4.12 eV, which is larger than that of the Au@LiNbO3 core/shell structure. This could be attributed to the increased particle size of Au@LN due to the presence of the core–shell structure, as shown in Fig. 6. On the other hand, increasing the energy gap with laser fluence is due to the decrease in core–shell particle size.

Fig. 5
figure 5

(αhν)2 versus photon energy plot of LN and Au@LiNbO3 prepared at different laser fluences

Full size image
Fig. 6
figure 6

Particle size and energy gap of Au@LN core/shell as a function of laser fluence

Full size image

The TEM images of Au@LN core/shell nanoparticles prepared at various laser fluences are shown in Fig. 7. These images confirm the formation of spherical nanoparticles with core/shell morphology, since the Au core nanoparticle is surrounded by a LN shell and the size of the core–shell particle depends on the laser fluence. The particle size increases first from 15 to 23 nm as laser fluence increases from 1.3 to 1.6 J/cm2 and then decreases to 19 nm as laser fluence increases to 2.2 J/cm2, and it shows a wide range of particle distribution. This gives approval to the change in shift of the absorption edge and the change in energy gap value of the UV–Vis results. Figure 8 shows the magnified TEM images of the core/shell structure, which revealed the formation of a monodispersed core/shell with a very clear boundary between Au-core and LN-shell. The effect of laser fluence on the core size and shell thickness is shown in Fig. 9. As laser fluence increases, the core size and shell thickness increase up to the laser fluence of 2 J/cm2 and further increases in laser fluence result in decreasing the core size and shell thickness due to fragmentation process.

Fig. 7
figure 7

TEM images of Au@LN core/shell nanoparticles synthesized at various laser fluences

Full size image
Fig. 8
figure 8

Magnified TEM images of Au@LiNbO3 core–shell nanoparticles synthesized at a 1.3, b 1.6, c 2, and d 2.2/cm2

Full size image
Fig. 9
figure 9

Relationship between core size and shell thickness with laser fluence

Full size image

Figure 10 shows the SEM images of Au@LiNbO3 core/shell nanoparticles prepared at different laser fluences. These images reveal the formation of spherical nanoparticles and many, agglomerated and aggregated nanoparticles are observed.

Fig. 10
figure 10

SEM images of core/shell prepared at various laser fluences

Full size image

Figure 11 shows the Raman spectra of Au@LiNbO3 core–shell nanoparticles recorded in the range of 95–1000 cm−1. For all samples except the one prepared at 1.6 J/cm2, Fig. 10 confirms the presence of seven E modes and one A Raman mode indexed as E(TO), E4, E(6TO), E(LO6), E(LO7), (E-TO8), 4A1TO, and E(LO9). This result emphasizes that all observed phonon modes in Raman spectra for synthesized samples are consistent with the vibration modes for LiNbO3 single crystal reported by other workers [44, 45]. As laser fluence increased, the peak intensity increases due to increasing the concentration of the core–shell nanoparticles. A significant enhancement in the Raman intensity may be produced after employing Au nanoparticles to LN structures due to the presence of plasmonic Au NPs, which play an essential role in the enhancement of the intensity of Raman peaks. This can originate from the local fields associated with the excitation of surface plasmon resonances by the Raman source [46, 47]. The vibration modes of the Au@LiNbO3 core/shell samples as a function of laser fluence are listed in Table 1.

Fig. 11
figure 11

Raman spectra of Au@LiNbO3 synthesized at various laser fluences

Full size image
Table 1 Raman vibration modes of Au @LiNbO3 core/shell nanoparticles
Full size table

Figure 12 presents the zeta potential plots of the Au@LiNbO3 core/shell synthesized at various laser fluences. The value of the zeta potential gives an indication of the potential stability of the colloidal Au@LiNbO3 core/shell. As shown in Table 2, the best value of ZP was 32 mV for the sample prepared at a laser fluence of 2.2 J/cm2, which indicates that colloidal suspensions are stable and no agglomeration and/or aggregation can be obtained. The origin of particle agglomeration is the van der Waals force and high surface energy. The nanoparticles synthesized at 2.2 J/cm2 have the highest value of ZP due to their high concentration. Ions are adsorbed on the surface of the particles, affecting the value of the isoelectric point and, as a result, increasing particle dispersion.

Fig. 12
figure 12

Zeta potential plots of Au@LiNbO3 core/shell nanoparticles

Full size image
Table 2 ZP value as a function of laser fluence of Au @LiNbO3 core/shell nanoparticles
Full size table

Figure 13 illustrates the dark forward and reverse current–voltage characteristics of p-Au@LiNbO3 NPs/p-Si heterojunction at room temperature. It is clearly seen that the samples exhibited rectifying properties, indicating the formation of diode-like devices. The highest rectification factor was found for heterojunction prepared at 2 J/cm2. As is obvious, the forward current increases as the bias voltage increases due to decreasing the depletion region, and the turn-on voltage was found to depend on the laser fluence. For heterojunctions prepared at 1.3 J/cm2 and 1.6 J/cm2, the turn-on voltage is less than 2 V, while it is greater than 2 V for samples prepared at 2 and 2.2 J/cm2. This can be justified as a result of decreasing the electrical resistivity of the Au@LiNbO3 core–shell as well as the dominance of diffusion current. The reverse current increases slightly for all samples with increasing bias voltage, particularly at voltages greater than 2 V. The forward current of the heterojunction prepared at 2 and 2.2 J/cm2 is larger than the other fabricated heterojunctions due to the decreasing of the electrical resistivity of the nanoparticles.

Fig. 13
figure 13

Effect of laser fluence on the dark forward and reversed current of p-Au@LiNbO3/p-Si devices. Inset is the schematic illustration of Au@LiNbO3/Si heterojunction with electrical configuration

Full size image

The ideality factor β of the heterojunction could be extracted from diode equation [50]

$$\mathrm I={\mathrm I}_{\mathrm s}\mathrm e^\frac{\mathrm{qV}}{\mathrm\beta\mathrm K\mathrm T}$$
(1)

where K is the Boltzmann constant, q is the electron charge, and Is is the saturation current. The saturation current was determined from the Ln (If) versus Vf plot, where If is the forward current and Vf is the voltage at forward bias as shown in Fig. 14.

Fig. 14
figure 14

Semi logarithmic relationship of forward current–voltage characteristics of heterojunctions prepared at different laser fluences

Full size image

The ideality factors of the p-Au@LiNbO3/p-Si heterojunction devices fabricated at laser fluences of 1.3, 1.6, 2, and 2.2 J/cm2 were 6.5, 3.9, 2.0, and 5.3, respectively. The large value of the ideality factor of the heterojunction indicates deviation from the ideal diode, suggesting the presence of defects like trapping, series resistance, and surface states arising from agglomeration and aggregation of the nanoparticles [51, 52]. The value of the ideality factor indicates that the heterojunction prepared at 2 J/cm2 has the best junction characteristics. By using thermionic emission of current–voltage characteristics of the Au@LN/Si heterojunction, the barrier height Φ can be given by [53]:

$$\Phi = \frac{KT}{q} Ln \frac{A{A}^{*} {T}^{2}}{{I}_{s}}$$
(2)

where A is the area of the heterojunction and A* is the Richardson constant. The values of the barrier height of the Au@LN NPs/Si heterojunctions fabricated at laser fluences of 1.3, 1.6, 2, and 2.2 J/cm2 were 0.84, 0.83, 0.85, and 0.82 eV, respectively. The capacitance voltage (C-V) characteristics of the heterojunctions are depicted in Fig. 15.The linear relationship indicates that the fabricated hetero-junctions are abrupt in nature. The value of the built-in-potential Vbi can be calculated by extrapolating the linear part to C−2 = 0. The values of the built-in potential were 0.6, 0.5, 0.4, and 0.8 V for the heterojunctions prepared at 1.3, 1.6, 2, and 2.2/cm2, respectively. As shown in Fig. 15, the capacitance of all heterojunctions is inversely proportional to the reverse bias voltage. The reduction in the device capacitance with bias voltage resulted from the widening of the depletion layer [54]. This properly gives an indication of the behavior of the charge transition between the junction sides [55]. Increasing the capacitance of the junction with laser fluence can be attributed to increasing the hole concentration of the core–shell nanoparticles. The linearity characteristics of the photodetectors are shown in Fig. 16. The photocurrent is plotted versus light intensity. This figure gives information about the linear dynamic range (LDR). The LDR is given by the following equation:

$$LDR=20Log \frac{{P}_{max}}{NEP}$$
(3)

where Pmax is the maximum amount of light that illuminated the photodetector and NEP is the noise equivalent power. The values of LDR estimated from Fig. 16 of the fabricated photodetector fabricated at 1.6, 2, and 2.2 J/cm2 were 59, 152, and 61 dB, respectively.

Fig. 15
figure 15

a Variation of junction capacitance with reverse voltage and b 1/C2 versus voltage of p-Au@LiNbO3/p-Si heterojunctions fabricated at different laser fluences

Full size image
Fig. 16
figure 16

Variation of photocurrent with light intensity of p-Au@LiNbO3/p-Si heterojunctions fabricated at different laser fluences

Full size image

Figure 17 shows the effect of laser fluence on the spectral responsivity Rλ of the p-Au@LiNbO3/p-Si photodetectors measured at reverse bias of − 8 V. A clear peak of response was observed at 399 nm for all heterojunctions, which can be indexed to the absorption edge of the Au@LiNbO3 core–shell and a second peak appeared at 850 nm which attributed to the silicon substrate [56][57]. The maximum responsivity was 0.43 A/W at 400 nm for the photodetector fabricated with 2 J/cm2. Furthermore, a small peak appeared at 600 nm for the photodetector fabricated, which may be attributed to the SPR effect of the Au core. It is clearly noticed that It is clear that the photodetector’s responsivity is dependent on laser fluence, which can be attributed to the increasing shell thickness with laser fluence. Increasing the shell thickness improves the responsiveness in the visible region. On the other hand, an increase in the electrical resistivity of the nanoparticles leads to an extension of the depletion region toward the Au@LiNbO3 and, in turn, enhances the short wavelength responsivity. The specific detectivity of the photodetector was calculated from the following equation:

$${\mathrm{D}}^{*}= \frac{{\mathrm{A}}^{0.5}{\mathrm{ I}}_{\mathrm{ph}}}{\mathrm{P}\sqrt{2{\mathrm{eI}}_{\mathrm{d}} }}$$
(4)

where A is the sensitive area of the photodetector, P is the light power, e is the electron charge, Id is the dark current of the photodetector, and Iph is the photocurrent of the photodetector. The specific detectivity D* of the photodetectors was determined as a function of laser fluence, as shown in Fig. 18. The maximum D* was 1.32 × 1012 and 1.08 × 1012 Jones at 380 nm and 850 nm, respectively, for Au @ Au@LN/p-Si photodetector fabricated at 2 J/cm2. The presence of Au nanoparticles is critical in reducing e–h recombination caused by the induction of a high electric field region and increasing photocurrent of the photodetector [58, 59]. The external quantum efficiency EQE of the photodetector can be defined is the ratio of incident number of photons to number of generated electrons. It can be determined from the following relationship:

Fig. 17
figure 17

Spectral responsivity of p-Au@LiNbO3/p-Si photodetectors fabricated at different laser fluences at bias of 8 V

Full size image
Fig. 18
figure 18

Specific detectivity of p-Au@LiNbO3/p-Si photodetectors at reverse bias voltage of − 8 V

Full size image
$$\mathrm{EQE}=\frac{1240{\mathrm{R}}_{\uplambda }}{\uplambda (\mathrm{nm})}$$
(5)

The value of EQE at 400 nm for the photodetectors prepared at laser fluences of 1.3, 1.6, 2, and 2.2 J/cm2 was estimated from Eq. (5) and found to be 0.33, 0.48, 1.3, and 1.13, respectively, at a bias voltage of − 8 V. As shown, the value of EQE is larger than unity, which can be explained as follows: the width of the depletion region widens as the reverse bias voltage is applied, suggesting that the device is fully depleted at this voltage, resulting in efficient photogenerated carrier collection [60]. Furthermore, the applied bias voltage improves charge injections and forms free carriers, which absorb light photons and increase quantum efficiency dramatically through the photomultiplication effect. The figures of merit of the fabricated photodetector have been compared with other heterojunction-based silicon photodetectors as depicted in Table 3.

Table 3 A comparison of the figures of merit of the fabricated photodetector with those for other heterojunction-based silicon photodetectors
Full size table

An energy band diagram has been constructed to understand the rectifying properties of the fabricated device, as shown in Fig. 19. The electron affinity of silicon is 4.05 eV, and its energy gap is 1.1 eV at room temperature, and the work function of Au is 5 eV. Crystalline Au@LN deposited using the LAL method is usually Li ion deficient. Its band gap at optimum laser fluence is found to be 3.8 eV. LN’s electron affinity is currently thought to be unaffected by impurity levels, with a value of 1.1 eV [65]. The band offsets of LN/Si heterojunction wee estimated and found to be ΔEC = χSi − χLN = 4.05 − 1.1 = 2.95 eV and Δ EV = χSi + EgSi − ( χLN + EgLN) = 4.05 + 1.12 − (1.1 + 3.8) = 0.27 eV. It is noteworthy that the electron jumps from Au to LN to substitute for the missing electron via the recombination process.

Fig. 19
figure 19

Energy band diagram of Au@LiNbO3/p-Si photodetector prepared at 2.2 J/cm2

Full size image

Conclusion

We have successfully synthesized an Au@LiNbO3 core/shell structure using pulsed laser ablation in ethanol without using catalyst. The variation of laser fluence affects the structural, optical, and electrical properties of Au@LiNbO3 core/shell. The presence of peaks related to LiNbO3 and Au in the XRD data revealed that the product has a polycrystalline structure and confirmed the formation of the core/shell structure. The TEM investigation confirmed the formation of core/shell nanoparticles with a spherical shape, with the gold core surrounded by a LiNbO3 shell. The laser fluence was found to increase shell thickness. The current–voltage properties of Au@LiNbO3/Si confirmed the formation of a good heterojunction with a rectification factor that depends on the laser fluence. The parameters of an isotype Au@LiNbO3/Si heterojunction photodetector as a function of laser fluence were investigated. The spectral responsivity results revealed that the fabricated photodetectors have two peaks of response, specifically located at 399 and 50 nm. The maximum responsivity was 0.43A/W at 399 nm for the photodetector fabricated with a laser fluence of 2 J/cm2. Increasing the responsivity in the visible region through the insertion of a buffer layer and selection the optimum preparation conditions is underway. The findings of this work confirm that the characteristics of core/shell synthesized in an ethanol medium are better than those obtained in distilled water due to the lower particle agglomeration.