Nowadays, ultraviolet photodetectors (UV photodetectors) play a crucial role in a variety of industries and applications where detection and measurement of UV radiation are required. These applications range from household to universe, and includes security cameras, medical and health care, agriculture and food control, missile guidance and target tracking^[1−5]. They offer high sensitivity and short response times, making them capable of detecting even trace levels of UV radiation. UV photodetectors are gaining popularity among researchers due to their wide range of practical applications and the rapidly expanding demand for photodetectors in the industry. Researchers have attempted to create UV photodetector utilizing populars materials include GaN^{[6, 7]}, ZnO^{[8, 9]}, CdSe^{[10, 11]}, PbS^[12], and WO₃^[13]. Among these materials, GaN is a promising candidate due to its outstanding characteristics such as wide band gap, low dark current, fast speed, long-term stability, and resistance to radiation^{[14, 15]}. The wide band gap in GaN, however, limits its use to the UV spectrum, prohibiting it from functioning as standalone functional layers in effective broadband photodetectors, which are critical for next generation image sensing applications^[16].

Metal nanoparticles are small metallic particles that typically range in size from 1 to 100 nm. These nanoparticles exhibit unique properties due to their small size and a high surface area-to-volume ratio. They come in a variety of shapes such as spheres, rods, or shells, and are widely used in optics-related fields. The integration of some metal nanoparticles such as silver (Ag), gold (Au), platinum (Pt) in GaN photodetectors can lead to enhance light absorption and improved photoresponse due to the excitation of surface plasmon^[17−19]. This can lead to higher efficiency, increased responsivity, and improved device performance in terms of detecting and converting light into electrical signals. Integration of Ag or Au nanoparticles (NPs) with GaN has been explored for enhancing performance of photodetector, however, the incorporating bimetallic Ag/Au NPs in UV photodetector has been rarely studied. Bimetallic NPs like Ag/Au can exhibit better qualities due to synergistic effects between the two metals. Both Ag and Au NPs exhibit plasmonic effects, which can enhance light absorption and photoconversion efficiency in the photodetectors^[20]. The combined impact of both metals can lead to improve overall performance. Furthermore, Au is recognized for its chemical stability, which can assist stabilize and protect the nanoparticles, resulting in a more stable and reliable photodetector over time^[21].

Herein, we present the high-performance photodetection device based on integration of bimetallic Ag/Au NPs/ GaN film heterostructures. The fabricated photodetector is examined for the key parameters including photocurrent, photoresponsivity, external quantum efficiency, and photoresponse speed, and the merits of NPs/GaN hybrid photodetector is presented.

The deposition process of GaN thin film was described in detail in our previous study^[22]. Sputtering was employed to fabricate GaN thin films on sapphire substrate using GaN target (99.99%). Sputtering was performed using a mixture of nitrogen and argon gases with a N₂/Ar ratio of 40 : 10 sccm and a base pressure of 1 × 10⁻⁶ mTorr. The average power and substrate temperature was kept at 75 W and 350 °C, respectively during deposition process. The GaN films were then annealed in N₂ ambient at 800 °C for 1 h to form crystalline films. To generate Ag−Au bimetallic nanoparticles on GaN films that had already been deposited, initial layers of thin Ag (8 nm thickness) were applied to the GaN films, followed by 5 nm thick Au films through thermal evaporation. Subsequently, the Ag−Au/GaN underwent annealing in a furnace for 1 h at a temperatures of 450, 550, and 650 °C (samples DV1, DV2, and DV3, respectively). Sample DV4 was deposited with 10 nm Au and then treated at temperature of 650 °C for 1 h. The samples were then cooled down for 24 h before being utilized for the creation of the electrodes. Each Ag−Au/GaN sample was coated with a pair of gold electrodes that were 100 nm thick, 3 mm long, and spaced 200 µm apart using thermal evaporation technique. Each sample was produced four times to test the reproducibility.

The Raman scattering investigation for GaN film was carried out using a LabRAM HR Evolution 532 nm laser. UV−Vis measurement was conducted on the transmittance spectra of the deposited GaN film. The crystallinity of Ag−Au NPs on GaN were examined using X-ray diffraction (XRD) in 2-theta mode. The morphologies of samples were analyzed using scanning electron microscopy, and the composition was evaluated using X-ray photoelectron spectroscopy with an Al K-alpha anode. The current−voltages curves were tested using a HP 4155 analyzer and UV light with a center wavelength of 365 nm. A monochromator with a 5 nm wavelength step was used to measure responsivity from DUV to visible.

Room temperature Raman spectra of GaN films grown on a planar sapphire is shown in Fig. 1(a). As shown in the figure, four peaks appear including sapphire E(g), sapphire A₁(g), GaN-E₂ (high), and GaN-A₁ (LO). E₂ (high) and A₁ (LO) are the active phonon modes of GaN, and the shift of E₂ (high) modes is utilized to calculate the strain or stress present in the GaN film. For free-standing GaN, the E₂ (high) mode should be observed at 567.6 cm⁻¹, however, in this investigation, E₂ (high) phonon mode peak was observed at 565.3 cm^−1[23]. This red shift of 2.3 cm⁻¹ implies that residual strain in the GaN grown film is compressive in nature. The strain generation is due to difference in lattice constant of GaN and sapphire substrate. The strain can be quantified by assuming a linear relationship between the E₂ (high) Raman shift Δw and the strain σxx as following^[24]:

1Δw=Kσxx.

With K is the linear proportionality factor between the shift in the strain and the E₂ (high) mode. The train calculated here in is 0.55 GPa in the GaN film grown by sputtering on sapphire substrate. Fig. 1(b) presents the transmittance spectra obtained at room temperature of bare GaN film. The high transmittance in the visible range (from 400 to 800 nm) is above 70% indicating that the as-grown GaN film is naturally transmittance. Moreover, the transmittance spectra shows a sharp steep transition edge at about 365 nm. The sharp cut-off edge could be seen at the ultra-violet region, suggesting that the GaN film has outstanding absorption at the UV region. To more clarify the UV absorption of GaN film, the optical band gap was determined by the following relation:

2(αhν)=β(hν−Eg)1/2,

where α = 1/dln(T). In this equation, α is the absorption coefficient, hν is the photon energy, β is a constant, Eg is the optical bandgap, d is thin film thickness, and T is the transmittance at the specified wavelength. Plotting the straight-line portion of the (αhν)² vs. hν yields the optical band gap of each film. The computed optical band gap of the as-grown GaN film obtained via sputtering process was 3.42 eV, which is consistent with the previous report^[25].

The structure of Ag−Au bimetallic NPs on GaN film was investigated by X-ray diffraction. Fig. 2(a) illustrates the X-ray diffraction patterns of the Ag−Au NPs/GaN treated at the various annealing temperatures. The 2θ values of Au−Ag bimetallic are very close to each other since they have similar lattice constants (JCPDS: 4-0783 and 4-0784). Two strong GaN peaks are observed at 40.23^o and 47.17^o that correspond to (200) and (102) planes^[26], respectively. The Au−Ag bimetallic direct to (111) and (200) orientation (JCPDS cards No. 04-0783) and a magnified image of plane (111) is shown in Fig. 2(b). The FWHM of the (111) peak has been used to evaluate the crystallite size (L) of the Ag−Au NPs using Scherrer equation:

3L=Kλβcosθ,

where β is the value of FWHM in radiant situated at 2θ of (111) plane, K is the shape factor which is typically considered as about 0.89, and λ = 0.15405 nm. Fig. 2(b) shows that increasing the annealing temperature from 450 to 650 °C significantly increases the FWHM 0.32 ^o to 0.43 ^o while decreasing the crystallite size from 33 to 26 nm, respectively. Specially, with the FWHM of just Au NPs was found to be greatest with the smallest crystallite size of approximately 19 nm. This demonstrated that the structural properties of grown metal atoms are altered by the thermal treatment and metal composition. At low annealing temperature, the nanostructure begins to mix and generate larger grain size and as temperature reaches 650 °C, the formation of Ag−Au bimetallic NPs completed with the smallest crystallite size^[27]. At the same annealing temperature of 650 °C, however, pure Au NPs (DV4) produced smaller crystallites than Ag−Au bimetallic NPs (DV3). This smaller size is attributed to the higher activation energy of simply Au atoms than Ag−Au atoms, which influenced the shorter diffusion length attained^[22].

The chemical composition of the fabricated Ag−Au/GaN was properly exploited by XPS. Fig. 3(a) shows the XPS survey of Ag−Au/GaN sample with the presence of N 1s, Ga 3d, Ag 3d, and Au 4f. Ga 3d and N 1s are derived from the as-grown GaN thin film, whereas Ag 3d and Au 4f come from the bimetallic nanoparticles. Fig. 3(b) illustrates the XPS spectrum of Ag (3d) peaks for the grown Ag−Au NPs and Au NPs treated with the various annealing temperatures. It is clearly shown in Fig. 3(b) that Ag (3d) peaks appear in all Ag−Au bimetallic NPs but not in DV4 where only deposited with Au NPs. When comparing the Ag (3d) binding energies of individual Ag NPs to those of the alloying Ag−Au, the latter shift to a lower value. The binding energies of the pure only Ag NPs: Ag (3d_3/2) and Ag (3d_5/2) are 368.9 and 374.8 eV, respectively^[28]. Ag (3d) peaks in Ag−Au bimetallic NPs showed a decreased in XPS intensity when annealing temperature was raised to 650 °C. One possible explanation could be that at low temperature, the thermal energy is insufficient to cause significant intermixing or alloying of the Ag and Au components. As a results, the Ag (3d) peaks remain distinct and the XPS intensity is higher at low temperature due to the presence of more pure Ag. Fig. 3(c) displays the XPS spectrum of Au (4f) peaks for deposited grown Ag−Au NPs and Au NPs after annealing treatment. The Au monometallic NPs shows the lower binding energies than Ag−Au NPs which can be assigned to the Au metallic state because there is no corresponding oxidization in pure Au NPs. This observation aligned with the earlier research findings^[28]. As in creasing temperature, the XPS binding energies Au (4f) of Ag−Au NPs shifts to the higher values. That can be attributed to the thermally-induced structural changes and oxidation when elevating temperatures.

Fig. 4(a) depicts the complete architecture of an Ag−Au NPs/GaN based-photodetector. Ag−Au bimetallic NPs were deposited on GaN film grown on sapphire substrate. To create the electrodes two Au films with the thickness of 100 nm were fabricated on GaN by thermal evaporation. The active distance between two Au electrodes is 200 µm which was determined by EDX mapping technique (Fig. 4(b)). Fig. 4(c) shows the photocurrent as a function of applied bias voltages of bare GaN and integration of NPs and GaN-based UV photodetectors excited by a LED with a wavelength center of 365 nm and a power of 0.87 mW/mm². It was discovered that the photocurrent of NPs/GaN devices was significantly enhanced than the bare GaN. In detail, the photocurrent of NPs/GaN devices is approximately 3 × 10³ times that of bare GaN-based device. The integration of Au or Ag−Au bimetallic NPs on the bare GaN revealed a considerable dominance in photocurrent improvement due to the LSPR effect^[29]. At bias voltage of 1 V, the photocurrent for DV1, DV2, DV3, and DV4 is 56.5, 66.5, 87.6, and 64.9 mA, respectively. As can be seen, when the annealing treatment on Ag−Au NPs increases, the photocurrent of corresponding device also increases. The photocurrent I_ph of 650 °C treated Ag−Au/GaN is about 1.55 times that of 450 °C at the same thickness. The rapid rise in I_ph with UV illumination at high temperature can be attributed to the complete formation of Ag−Au bimetallic NPs at 650 °C that was demonstrated in our previous study^[27]. Another possible factor is the enhanced crystallinity, as seen by above XRD results. The better crystallinity of the NPs in turn can improve their electronic properties resulting in more effective charge transfer between the NPs and GaN, and hence larger photocurrent. Compared to pure Au NPs/GaN-based device, the Ag−Au NPs annealed at 650 °C also exhibited a higher photocurrent. The larger photocurrent was attributed to significantly better light absorption and hot electron transfer in the Ag−Au NPs than pure Ag or Au NPs, leading to the rise of photocurrent generation^[29]. Fig. 4(d) presents the ON/OFF ratio of bare GaN-photodetector and the NPs/GaN-photodetectors at bias voltage of 1 V, UV illumination with wavelength center of 365 nm. The bare GaN-based device exhibits the maximum on/off ratio of 988. The ON/OFF ratio values of DV1, DV2, DV3, and DV4 are 496, 549, 903, and 705, respectively. The higher obtained ON/OFF ratio of DV3 than the others could originate from the higher photocurrent and the lower dark current of DV3 (treated at 650 °C). Despite of much lower photocurrent, the bare GaN-based photodetector still shows high ON/OFF ration due to its low dark current. The lower dark current is the key factor to the low noise in the photodetector. Photocurrent and ON/OFF ratio findings indicate that the Ag−Au NPs/GaN which annealed at 650 °C (DV3) is the best candidate showing the excellent optoelectronic characteristics.

Other key parameters used to justify the quality of a photodetector include the response speed, photoresponsivity (R), external quantum efficiency (EQE), and detectivity (D^*). Figs. 5(a) and 5(b) show current on the ON/OFF state as the function of time of photodetector on bare GaN and with Ag−Au NPs annealed at 650 °C, respectively. The found rise time response τd reduced from 0.52 to 0.12 s with the inclusion of Ag−Au NPs. The persistence photoconductivity (PPC) effect was more pronounced in device with Ag−Au compared to those without, with a decay time of approximately 1.67 s. The presence of Ag−Au NPs led to inter-diffusion, as confirmed by XPS data, causing the PPC effect due to unintended doping. Fig. 5(c) illustrates the responsivity curve as function of incident wavelength of DV3 at a bias voltage of 1 V. This photodetector respond selectively to a specified range of UV range (λ = 230−400 nm), whereas the peak center of R is at 340 nm. The greatest photoresponsivity achieved at λ = 340 nm is 98.5 A/W which is substantially greater than previous studies on UV photodetectors^{[30, 31]}. Furthermore, this measured photoresponsivity value was significantly better than that found in bare GaN-based photodetector with the peak center of R of 1.67 × 10⁻³ A/W. In addition, there is a noticeable broadband from 500 to 650 with a shoulder peak at about 600 nm appearing in photoresponse curve. The inclusion of Ag−Au nanoparticles into UV photodetectors causes plasmon resonance effects, in which the nanoparticles resonate at visible wavelength. This can further improve the absorption of light in the visible band, contributing to the broadening of the photoresponsivity curves. Another hypothesis could be from the defects in the GaN film such as substitutional oxygen (O_N) and gallium vacancy (V_Ga)^[32]. The external quantum efficiency (EQE) is derived from the responsivity as the following relation:

4EQE=Rλhceλ.

The maximum EQE was similarly reported at approximately 340 nm, with a value of 369%. Although the sub-broaden peak was observed in photoresponse curve, the EQE only responded to the selective region of UV light from 250 to 400 nm. Therefore, the Ag−Au NPs/GaN structure is perfectly suitable to fabricate UV photodetector. The plasmonic NPs on the GaN can greatly enhance the photocurrent, photoresponsivity, and EQE of Ag−Au NPs-based devices.

The improvement of photocurrent in UV photodetector with plasmonic NPs can be explained by examining how incident photons interact with the NPs and considering energy band theory, as depicted in Fig. 6. When exposed to UV light, GaN absorbs energy from photons, leading to the creation of electron−hole pairs, which in turn leads to the generation of photocurrent in bare GaN-based device. While the Ag and Au NPs could cause the activation of electrons, leading to the production of high-energy hot carriers through nonradiative plasmonic decay^[33]. The heighted energy levels of the electron−hole pairs generated during surface plasmon decay in metal nanoparticles greatly exceed those of carriers close to the Fermi level. When the metal nanoparticles and GaN come into contact, the energy band aligning in a way that matches the Fermi energy because the Ag and Au metals have a higher work function (4.26−4.7 eV and 5.1−5.47 eV, respectively) compared to n-type GaN (4.2 eV)^[29]. That time, the collective oscillation of electrons in metallic NPs can boost their performance, as excited hot electrons can easily transfer to the conduction band of GaN that significantly increase the device’s photocurrent and responsivity.

In summary, we successfully constructed the UV photodetectors by incorporating Ag and Au NPs onto GaN films grown using sputtering technique. When the annealing temperature for Ag−Au alloys is increased, the crystallite size increased indicating to the increase of crystallinity of Ag−Au NPs. By integrating both Ag and Au NPs on GaN, the photoresponse improves considerably than bare GaN and Au NPs/ GaN. The best performance was observed in Ag−Au NPs/GaN that was annealed at 650 °C with the photoresponse of 98.5 A/W. The findings highlight the importance of the NPs in promoting device performance. This presents a design rule for a cost-effective product solution for high-performance UV photodetectors.