© The Authors 2025

1. Introduction

When two quasars with a small separation are observed, there are often three different physical scenarios involved. The most common scenario is projected quasars, where quasars coincidentally appear very close to each other along the line of sight, but are actually at vastly different redshifts. A rarer scenario is dual quasars, which are at similar redshifts and physically interacting. The rarest scenario is lensed quasars, where the light from a single quasar is bent, resulting in two images of the same quasar. In that case, an intervening lensing galaxy is usually expected.

The three categories mentioned above each play different roles in the investigation of the universe. Gravitationally lensed quasars not only enable investigations that have typically been addressed by galaxy-galaxy lenses in the past, such as constraining the equation of state of dark energy and investigating galaxy evolution (Oguri et al. 2014; Suyu et al. 2014; Shu et al. 2015; Sonnenfeld & Cautun 2021; Filipp et al. 2023), but also provide a unique avenue for measuring the Hubble constant (Shajib et al. 2018; Liao et al. 2019; Wong et al. 2020; Li et al. 2021). Moreover, they offer information on the structures of active galactic nuclei (AGN, Anguita et al. 2008; Sluse et al. 2011; Motta et al. 2012; Guerras et al. 2013; Braibant et al. 2014; Fian et al. 2021; Hutsemékers & Sluse 2021). Dual quasars, which refer to quasar pairs separated by 1 pc to 100 kpc (De Rosa et al. 2019), have considerable significance in elucidating the growth and evolution of binary supermassive black holes (BSMBHs, Roedig et al. 2014; Romero et al. 2016) and illuminating the processes of galactic-scale merging (Boylan-Kolchin et al. 2008; Martin et al. 2018). Projected quasars can be used to study the circumgalactic medium (CGM, Cai et al. 2019) properties of quasar host galaxies (Findlay et al. 2018; Chen et al. 2023).

The discovery of those objects involves two key stages: candidate selection followed by spectroscopy confirmation. The first often rely on imaging surveys such as Dark Energy Spectroscopic Instrument Legacy Imaging Surveys (DESI-LS, Dey et al. 2019), Kilo Degree Survey (KiDS, de Jong et al. 2019), and Panoramic Survey Telescope and Rapid Response System (Pan-STARRS, Flewelling et al. 2020), or astrometric satellites such as Gaia (Gaia Collaboration 2018). The second relies on spectroscopic surveys, such as Sloan Digital Sky Survey (SDSS, Blanton et al. 2017) and Dark Energy Spectroscopic Instrument (DESI, DESI Collaboration 2016), as well as targeted follow-up observations (e.g., Lemon et al. 2022; Dux et al. 2023, 2024).

Based on a random forest (RF) selection and DESI-LS data, 620 new candidate lensed quasars have been collected in He & Li (2022), He et al. (2023). In this study, we selected ten high-priority observable candidates from He et al. (2023) and obtained their spectra using the DBSP/P200 at Palomar Observatory. Most of these targets exhibit separations smaller than 2.4″. Prior to our observations, we checked these targets in recently available spectral datasets, including SDSS DR16 and DESI EDR, to verify whether they had been spectroscopically observed in recent surveys. For confirmed lensed quasars, we performed light and lens modelling to reveal their properties. For confirmed dual quasars, we calculated the velocity differences, determined the projected separations, and discussed the CGM features.

The structure of this paper is as follows. Section 2 details the observational aspects, including a review of the target selection process, spectral data reduction procedures, and the specifics of spectroscopic follow-up. Sections 3 and 4 present the confirmation results of P200/DBSP observations and publicly available datasets, respectively. Finally, Section 5 provides a comprehensive discussion and summary of our findings. In this paper, we assume a fiducial cosmological model with Ω_m = 0.26, Ω_DE = 0.74, h = 0.72, w₀ = −1, and w_a = 0. Unless otherwise stated, all magnitudes quoted in this paper are in the AB system.

2. Observations

In Section 2.1, we briefly review the target selection process, with further details available in He & Li (2022), He et al. (2023). Section 2.2 introduces the spectroscopic equipment setups and observing conditions. Finally, Section 2.3 provides a brief overview of the spectral data reduction process.

2.1. Target selection

The first step in our approach involves employing a RF classifier to identify quasar candidates from DESI Legacy Surveys (DESI-LS) photometry catalogue. Our training set comprises 651 073 positive and 1 227 172 negative samples, incorporating photometric data from both DESI-LS and Wide-field Infrared Survey Explorer (WISE). The labels for these samples were derived from Sloan Digital Sky Survey (SDSS) and Set of Identifications, Measurements, and Bibliography for Astronomical Data (SIMBAD).

We applied the trained RF model to point-like sources from DESI-LS Data Release 9. The performance of the classifier, evaluated using a test set, demonstrated a recall of ∼99% at a purity of ∼30%. Through this process, we successfully identified approximately 24 million quasar candidates from a pool of ∼0.42 billion point-like sources. In the second phase, we designed a quasar group finder algorithm based on spatial coordinates to identify quasars in close proximity to each other. This algorithm initially generated ∼560 000 quasar groups from the ∼24 million quasar candidates. We then refined these groups by analysing the similarity of colours among group members and their likelihood of being quasars. This refinement process reduced the number of quasar groups from ∼560 000 to ∼140 000.

The final stage involved a visual inspection to select candidate strongly lensed quasars based on the spatial configuration of group members. During this process, we classified candidates into categories A, B, and C, with grade A representing the most promising candidates. This examination yielded 620 new lensed quasar candidates, consisting of 101 grade-A, 214 grade-B, and 305 grade-C candidates. Following the naming procedure used in He et al. (2023), we refer to the catalogue as H22.

2.2. Spectroscopic follow-up

The selection of targets followed a two-step process. The criteria of the first step were:

• Magnitudes of quasars brighter than 21.5 in the r band;

• No prior spectroscopic observations reported;

• Classified as grade A or B in H22;

• Observable from the Palomar observatory on 15–16 October 2023.

Subsequently, two additional conditions were considered, with one of them being sufficient:

• An extended component identified between two point-like sources in the residuals of Legacy Survey Imaging;

• Also flagged as candidates in Dawes et al. (2023).

After applying these criteria, 21 targets were selected, of which 10 were observed due to weather constraints. Long-slit spectroscopic follow-up was carried out for those ten candidates, using the Double Spectrograph (DBSP) equipped on the P200 on the nights of 15–16 Oct. 2023 (P.I. Zizhao He). Their magnitudes and separations can be found in Table 1.

The brightest magnitude is r = 18.45 and the faintest is r = 21.14, with a median value of r = 19.60. The largest separation is 4$\stackrel{″}{.}$65, while the minimum is 0$\stackrel{″}{.}$83, with a median value of 1$\stackrel{″}{.}$78. The composite image of the grz bands can be found in Figure 1. The standard dichroic D-55 was used to split light into blue and red ends. The 300 lines/mm grating blazed at 3990 Å was chosen for blue arms, while the 316 lines/mm grating blazed at 7150 Å for red arms. This provides a dispersion of 2.108 Å/pixel and 1.535 Å/pixel for the blue and red arms (Oke & Gunn 1982).

Fig. 1. DESI-LS colour images of the 10 targets observed by DBSP/P200 during Oct 15–16, 2023. The cutout size of each image is 12″ × 12″ and north is up, while east is left. These cutout images are obtained from Legacy Survey Imaging DR10 in grz bands and re-plotted in this work using HumVI (Marshall et al. 2016).

Seeing conditions generally ranged between 0.9″ and 1.5″, and a 1.5″ wide slit was utilized. For each exposure, the slit was aligned with two point sources, ensuring the sources were as close to the centre of the slit as possible. This approach ensured simultaneous and efficient acquisition of spectra for both sources.

Table 1 presents the key observational characteristics of the targets, including coordinates, exposure times, image separations, LS-DR10 r-band magnitudes, and determined redshifts. In addition, the table provides the results of the spectroscopic identifications.

2.3. Spectral data reduction

The raw spectra were reduced using a PYTHON¹ code developed by us. The code utilizes several well-known Python modules, including Astropy (Astropy Collaboration 2022), NumPy (Harris et al. 2020), and Pandas (Wes McKinney et al. 2010). We implemented a standard data reduction procedure that includes bias subtraction, flat-fielding, cosmic ray rejection (Astroscrappy, McCully et al. 2018), 1D spectrum extraction, and wavelength and flux calibrations.

Wavelength calibration at the blue and red ends was performed using FeAr and HeNeAr blended arc lamps, respectively. Flux calibration for the blue end was carried out with standard stars Feige 15 and Feige 34, while the red end was calibrated using Feige 34. Furthermore, all reduced spectra, along with those from SDSS DR16 and DESI EDR, were smoothed using Gaussian filtering to reduce noise and enhance spectral features (see Figs. 2, 6, 8 and 9).

Fig. 2. DBSP/P200 spectra and their corresponding cutout images of the two confirmed lensed quasars. The slit placement is labelled using the yellow lines.

3. P200-DBSP confirmation results

The observation results of ten targets using P200 in Palomar Observatory are presented here. Combing spectra and Legacy Survey Imaging, they were categorised into different classes, including lensed quasars (Section 3.1), likely lensed quasars (Section 3.2), dual quasars (Section 3.3), and projected quasars (Section 3.4). Spectroscopic analysis reveals that J0621+5227 is a star-star system composed of K-type stars. As this stellar system is out of the scope of our primary scientific objectives, we do not discuss it further in the subsequent analysis.

3.1. Confirmed lensed quasars

3.1.1. J0746+1344

Two point sources can be distinctly identified in the Legacy Survey Imaging, with r-band magnitudes of 18.45 and 20.04, respectively, separated by 2$\stackrel{″}{.}$44. The spectroscopic analysis reveals two quasars at a redshift of z = 3.105, exhibiting similar continuum and broad emission line profiles. This resemblance suggests that this is a gravitationally lensed system. In the light modelling procedure, the grz-band images are well fitted by two point spread functions (PSFs) that represent the lensed quasar images, along with one Sérsic profile that represents the lensing galaxy. We refer to the top line of Figure 3 for the image reconstruction results and Appendix A for the details on light modelling.

Fig. 3. Image decomposition results in grz-band for our confirmed lensed quasars. First: J0746+1344. Second: J2121−0826. First column (A): Original data. Second (B): Light model for two images. Third (C): A-B; Fourth (D): Light model for best-fitted lensing galaxy assuming Sérsic profile. Fifth (E): A-B-D.

Based on the spectral similarities between the two quasar images, the detection of an extended component consistent with a lensing galaxy and the successful fitting of a two-PSF plus Sérsic model fit, we have classified this system as a gravitational lens. The light distribution of lensing galaxy is fitted by a single Sérsic profile, with the results summarised at Table 2. The results of multiple images are summarised in Table 3.

The subsequent lens modelling was performed using a singular isothermal ellipsoid (SIE, Kormann et al. 1994) model. We input the position of observed quasars, while assuming the SIE centre is same as light profile of lensing galaxy. The SIE result is summarised at Table 4. In addition, the time delay is estimated as a function of lens redshift, with the result given in Figure 5. The magnification (μ) of two images is also given in Table 3. The detailed methodologies used in the mass modelling and time delay estimation can be found at Appendix A.

3.1.2. J2121−0826

In Legacy Survey Imaging, the r-band magnitudes of the two images are 19.63 and 20.55, respectively, which are separated by 1$\stackrel{″}{.}$34. A spectroscopic analysis of the DBSP data reveals a blended quasar spectrum at a redshift of z = 2.395. The system exhibits two distinct detections in Gaia data, suggesting the presence of multiple images. In the Legacy Survey Imaging residual, a red component becomes visible after subtracting two PSFs. The grz-band composite image is well modelled by simultaneously fitting a Sérsic profile positioned between the two PSFs (see the bottom line of Figure 3) consistent with the presence of a lensing galaxy. Based on this evidence (i.e. the blended high-redshift quasar spectrum, multiple Gaia detections, and the presence of a red component consistent with a lensing galaxy) we classified this system as a gravitational lens.

The Sérsic parameters are presented in Table 2. The PSF parameters can be found in Table 3. The SIE parameters are detailed in Table 4. Additionally, the time delay as a function of lens redshift is shown in Table 5.

Fig. 4. Modelled caustic and critical curves, along with the positions of images for the two confirmed lenses shown in Figure 2, are compared with the observed positions of multiple images. The plots are centred around the lens light centre, which is determined through light modelling and provided in Table 4.

Fig. 5. Time delay as a function of zd of two confirmed lensed quasars.

3.2. Likely lensed quasars

3.2.1. J0058+2524

According to Legacy Survey Imaging, the r-band magnitudes of two images are 18.94 and 19.71 respectively; two quasars are separated by 0$\stackrel{″}{.}$83. We have exposed this for 1600 s. A spectroscopic analysis of the DBSP data reveals a blended quasar spectrum at a redshift of z = 2.580, accompanied by two distinct detections in Gaia data. When modelled using a combination of two PSFs and one Sérsic profile, a potential lensing galaxy has been detected between the two PSFs, albeit at a low significance level of ∼1σ.

The presence of a blended quasar spectrum at z = 2.58 and multiple Gaia detections are promising indicators of a gravitational lensing system. However, the low detection significance of the lensing galaxy introduces uncertainty into this classification. Given the ambiguity regarding the lensing galaxy detection, we conclude that deeper, higher resolution imaging is necessary to conclusively confirm or refute this gravitational lens candidate.

3.2.2. J0122−1344

From Legacy Survey Imaging, two point-like components with different colours are visible. In the residual imaging of Legacy Survey Imaging, an extended yellow component is clearly seen. We performed light modelling on this system, and the results are summarised in Table 5. Here, we can identify three key observational facts. Firstly, typical quasar broad-line emissions, such as C_IV and Mg_II, can be identified from the blended spectrum, indicating the presence of at least one quasar. However, the continuum is very different from typical AGN continuum. Secondly, an extended galaxy is clearly visible from Sérsic + two PSF light modelling, with results indicating the Sérsic galaxy with magnitudes of m_r = 19.48 and m_z = 20.51. Thirdly, Figure 7 presents the g–r vs r–z plot for two images. Both the magnitudes from DESI-LS and our modelling are shown. The results from DESI-LS suggest image B is more plausibly a star, while image A is more plausibly a quasars, whereas our light modelling indicates that both images are quasars. These observations can be explained by two scenarios, detailed below.

• Lensing Scenario: Images A and B are doubly-imaged quasars. A bright lensing galaxy significantly contributes to the spectrum, causing the continuum to deviate from a typical quasar spectrum. The colour difference between the two images can be attributed to the blended light from the lensing galaxy.

• Quasar-Galaxy-Star Scenario: The two images represent a quasar-star pair, with a galaxy coincidentally positioned between them. According to Figure 7, the lensing scenario is more plausible because the colour of two images are differ from typical star locations.

To distinguish between these scenarios, improved spectroscopic observations that can capture two distinct spectra from the two point sources are necessary.

Fig. 6. DBSP/P200 spectra and their corresponding cutout images of the three likely lensed quasars. The slit placement is labelled using the yellow lines.

Fig. 7. Colour-colour plot for J0122−1344 is displayed, with blue points indicating stars. The yellow and purple pentagrams represent J0122−1344A based on Legacy Survey Imaging photometry. Furthermore, the yellow and purple triangles correspond to J0122−1344A and J0122−1344B, respectively, as determined by the light modelling results from this study, which utilises two PSFs, along with one Sérsic profile (refer to Appendix A).

3.2.3. J1556+7925

The DBSP data reveal a blended quasar spectrum at z = 1.507 and two distinct Gaia detections support the hypothesis that this system is a lensed quasar. However, no lensing galaxy is detectable after subtracting two PSFs in the Legacy Survey imaging. We believe deeper imaging is necessary to confirm this candidate.

3.3. Dual quasars

3.3.1. J1759+3459

A spectroscopic analysis reveals the presence of two distinct quasars at redshifts z = 1.970 and z = 1.982, respectively. The difference in redshifts effectively rules out the possibility of gravitational lensing. We classified this system as a binary quasar pair with an angular separation of Δθ = 2.3″, corresponding to a projected physical separation of 19.57 kpc at z = 1.970. The line-of-sight velocity difference between the two quasars is Δv = 1208 ± 14 km/s.

3.3.2. J1929+6009

A spectroscopic analysis reveals the presence of two distinct quasars at a redshift of z = 3.277, separated by an angular distance of Δθ = 1.34″. The possibility of gravitational lensing can be confidently ruled out due to the significant differences observed in the continuum spectra of the two quasars. These spectral variations may be attributed to differences in the accretion disk properties, dust extinction, or host galaxy contributions between the two quasars. We classified this system as a dual quasar pair with a projected physical separation of 13.4 kpc at the observed redshift, featuring a velocity difference of less than 10 km/s.

3.4. Projected quasars

3.4.1. J0422+0047

Initially thought to be a gravitationally lensed quasar system, with the central yellow component acting as the lensing galaxy, this intriguing configuration has been revealed to be more complex. The spectroscopy of the two quasar objects and the central component has definitively ruled out the lensing scenario due to the significant quasar redshift differences observed.

The large Δv of 11633 ± 14 km/s between the quasars leads us to classify this system as a chance alignment of two projected quasars with an intervening galaxy. Despite not being a gravitational lens, this serendipitous alignment enables us to investigate the CGM properties of both the quasar host galaxies and the intervening galaxy.

3.4.2. J2303+0814

Spectroscopic analysis reveals two distinct quasars with significantly different redshifts of z = 0.948 and z = 0.928, respectively. The substantial difference in their continuum spectra, coupled with the redshift difference, definitively rules out the possibility of gravitational lensing. The redshift difference between the two quasars is Δz = 0.020, which translates to a line-of-sight velocity difference of Δv = 8583 ± 22 km/s. This large velocity separation far exceeds what would be expected for physically associated quasars, further supporting our classification.

4. Inspection on available data-sets

We conducted a cross-match between H22 systems and several spectroscopic datasets with radius equal to 1″ (DESI-EDR and SDSS-DR16). Our analysis revealed new dual quasars (Section 4.1) and projected quasars (Section 4.2). To the best of our knowledge, these systems have not been previously discussed in the literature.

Table 6 summaries the key information for these quasar systems, including their right ascension (RA), declination (Dec), data source, angular separation, redshifts, and associated redshift errors. Figure A.1 presents the spectra of the dual quasars, along with their velocity differences (Δv) and projected separations.

4.1. Dual quasars

4.1.1. J0043+0424

A spectroscopic analysis from BOSS reveals that two of the systems exhibit characteristic broad emission lines typical of AGN. Notably, a self-absorption feature is observed in the C_IV emission line of quasar A, while it is absent in quasar B. This distinct spectral difference strongly suggests that we are observing a dual quasar system at z = 2.458 and z = 2.445, separated by Δθ = 2.78″ (rather than a gravitationally lensed quasar). On the other hand, the small value of Δv = 1116 ± 25 km/s suggests they are not projected quasars. Thus, we classified it as a dual quasar.

4.1.2. J0154−0048

The spectra from BOSS exhibit typical AGN features, including prominent broad emission lines such as C_III], with slightly different redshifts between the two quasars. The redshift difference corresponding to a Δv of 680 ± 141 km/s. According to studies by Oguri & Marshall (2010) and Cao et al. (2024), the z-band magnitude (m_z) for such lensing systems typically ranges from 14 to 21, which is brighter than the detection limit of the Legacy Imaging Surveys. Given this, we would expect to detect a deflector if this were a gravitational lens system. However, no such deflector has been observed. Consequently, we classified this system as a dual quasar with a separation of Δθ = 2.3″ or d = 19.48 kpc at z = 1.4817.

4.1.3. J1424+3439

A spectroscopic analysis from DESI reveals two sources exhibiting quasar characteristics at slightly different redshifts, which can be transferred to a Δv of 87 ± 39 km/s. Furthermore, no evidence of a gravitational lens deflector is detected in the residual images from the Legacy Survey. Consequently, we classified this system as a dual quasar pair with a projected separation of Δθ = 2.07″ or d = 17.31 kpc at z = 1.4716.

4.1.4. J1643+3156

A spectroscopic analysis from SDSS and BOSS confirms that components A and B are both quasars at approximately z = 0.586. According to the studies of Oguri & Marshall (2010) and Cao et al. (2024), the m_z value of a potential lensing galaxy for a system with z_s = 0.586 should be greater (brighter) than 17. Such a galaxy would be unambiguously detectable in the Legacy Imaging Surveys. Given the absence of any visible deflector galaxy in the LIS data, we can rule out the lensing scenario for this system. Therefore, we classified this system as a dual quasar pair with a projected separation of Δθ = 2.33″ or d = 15.44 kpc at z = 0.5853. The velocity difference is 67 ± 12 km/s.

4.2. Projected quasars

4.2.1. J0134+3308

The spectra under examination come from BOSS. This particular case presents an intriguing projected quasar pair with an angular separation of 3.56″. While the two quasars exhibit remarkably similar colours, their redshifts differ significantly, with one at z = 0.7064 and the other at z = 1.1359. Interestingly, both quasars show evidence of their host galaxies in the spectral residuals after subtracting the quasar contribution.

4.2.2. J0843+4733

Broad-line emissions such as C_IV, C_III, and Mg_II have been identified for both quasars. The spectrum of quasar A is of particular interest, as it displays a significant C_III absorption feature at ∼4000 Å. This absorption is potentially due to the presence of quasar B at a lower redshift. This type of configuration presents a valuable opportunity for investigating the CGM of the quasar host galaxy.

4.2.3. J1201−0117

The DESI-EDR spectra reveal two quasars at markedly different redshifts: one at z = 2.5797 (quasar A) and another at z = 1.0653 (quasar B). The emission of 9400 Å does not appear to be typical for AGN; thus, it probably comes from the atmosphere.

4.2.4. J1246+5030

The SDSS spectra reveal two quasars at significantly different redshifts: one at z = 2.7307 (quasar A) and another at z = 2.1129 (quasar B). This substantial redshift difference creates an intriguing cosmic alignment that offers a unique opportunity for studying quasar absorption systems. In the spectrum of the higher-redshift quasar A, a significant C_IV absorption feature is clearly seen. This absorption is particularly interesting as it may be due to the presence of quasar B, which lies at a lower redshift along the same line of sight.

4.2.5. J1351+5224

BOSS spectroscopic observations reveal two quasars with significantly different redshifts (z = 3.2027 and z = 0.9747). The spectrum of quasar A exhibits a prominent, broad Lyman-α emission line accompanied by the characteristic Lyman-α forest.

4.2.6. J1537+3649

BOSS spectroscopic observations reveal two quasars with distinctly different redshifts (z = 1.3850 and z = 1.0629). Based on this significant redshift discrepancy, we classified this system as a projected quasar pair separated by Δθ = 2.53″.

4.2.7. J1603+5449

DESI-EDR spectroscopic observations reveal two quasars with distinctly different redshifts (z = 1.5171 and z = 0.3451). Based on this significant redshift discrepancy, we classified this system as a projected quasar pair separated by Δθ = 2.24″.

4.2.8. J1655+3408

Spectroscopic analysis from the DESI-EDR confirms that quasars A and B are distinct objects. They are separated by or a projected separation of d = 28.25 kpc at z = 1.7316. The individual redshifts of the quasars (z = 1.7316 and z = 1.6931 respectively) yield a velocity difference of 4261 ± 37 km/s, which exceeds conventional boundary for dual quasars (2000 km/s, Hennawi et al. 2010). Consequently, we classified this system as projected quasar pair.

4.2.9. J1758+6654

DESI-EDR spectroscopic data reveals a pair of quasars with markedly different redshifts (z = 1.9550 and z = 2.9151). The spectrum of the higher-redshift quasar (B) exhibits a significant Mg_II absorption feature at ∼8400 Å. Intriguingly, this absorption may be attributed to the presence of the lower-redshift quasar (A), suggesting the light emitted by quasar B is absorbed by A, despite their considerable redshift difference.

5. Summary and discussion

In this study, we conducted follow-up spectroscopic observations of lensed-quasar candidates from the previously compiled H22 catalogue (He et al. 2023). These observations took place on 15–16 October 2023, at the Palomar Observatory in California, utilising the P200/DBSP instrument. Additionally, we cross-checked all H22 candidates with publicly available spectroscopic surveys, including DESI-EDR and SDSS. Through the combined efforts of Legacy Survey Imaging, DBSP/P200, and existing spectroscopic surveys, we successfully confirmed two lensed quasars, six dual quasars, and eleven projected quasar pairs. Our results are summarised below.

• Strongly lensed quasars: J0746+1344 and J2121−0826 are two lensed quasars at z_s = 3.105 and z_s = 2.395. Assuming SIE mass model, the θ_E of the lensing system are 1$\stackrel{″}{.}$208 and 0$\stackrel{″}{.}$749, respectively. Both lenses were confirmed through DBSP spectroscopy.

• Dual quasars: We identified six dual quasars, with two confirmed by DBSP spectra and the remainder from public datasets. The mean redshift of these pairs is 1.96, with J1929+6009 at z = 3.28 being the highest redshift pair and J1643+3156 at z = 0.59 the lowest. The projected physical separations range from 15.44 to 22.54 kpc.

• Projected quasar pairs: Eleven projected quasar pairs were confirmed: two through DBSP spectra and the rest from public datasets. The projected physical separations of these pairs, ranging from 10.96 to 39.07 kpc, make them suitable for studying the CGM of the lower redshift quasars (Lau et al. 2018; Chen et al. 2023).

• Potential lensed quasars: Based on DBSP spectra and Legacy Survey Imaging, we identified three potential lensed quasars. These candidates require deeper imaging or resolved spectra for each image to conclusively determine their lensing nature.

One of the confirmed lensed quasars, J0746+1344, exhibits a strong flux anomaly. The lensing galaxy is located next to the brightest image (image B), which is unusual since the opposite configuration is typically observed. According to Legacy Survey Imaging, over a two-year baseline, the magnitude of image B is more than one magnitude brighter than that of image A in the grz bands. This is likely due to an ongoing microlensing effect.

We note that J0746+1344 lies within the WFST wide-file survey footprint (Chen et al. 2023), making it ideal candidate for time delay measurement, provided its time delay is less than 90 days. Both of the confirmed lensed quasars can be monitored by the Muztagh-Ata 1.9-meter Synergy Telescope (MOST, see e.g., Zhu et al. 2023) in the future. Projected quasar pairs serve as exceptional cosmic probes, playing a crucial role in studying the CGM of quasar hosts and their environments. Among our discoveries, J0422+0047 stands out as a particularly intriguing system. It offers the unique opportunity to simultaneously explore the CGM in both the outskirts of the quasar host and a coincidentally overlapped foreground galaxy. This rare configuration provides valuable insights into the diverse environments of the CGM, highlighting differences between AGNs and quiescent elliptical galaxies. For the potential lensed quasars, deeper and higher resolution imaging can be anticipated from future missions such as Euclid (Euclid Collaboration: Mellier et al. 2025), Chinese Space Station Telescope (CSST; Cao et al. 2018; Zhan 2021), and Roman (Eifler et al. 2021). If confirmed, although their individual light curves cannot be resolved or separately recovered by WFST or MOST, time-delay measurements can still be conducted using blended light curves (Shu et al. 2022).

In addition, we would like to emphasise an intriguing dual quasar discovered by P200, J1929+6009 at z = 3.28. The redshift difference between the two components is less than 0.0001 and the projected separation is 19.28 kpc. In Hennawi et al. (2010), 24 dual quasars were identified within a redshift range of 3 to 4, with a mean projected separation of 267 kpc. This highlights the uniqueness of the J1929+6009 system: its projected separation of 19.28 kpc at z = 3.28 is notably small. Another interesting aspect is numerous absorptions spanning the range of 3500 Å to 5200 Å have been observed on the blue side of J1929+6009B, which are not visible in J1929+6009A, although they are only separated by Δθ = 1.76″ and at the same redshift. As illustrated in Figure 10, the most prominent feature appears to be a Ly_β absorption at z = 2.797. However, no counterpart can be found a such a redshift in the public available spectra datasets.

Fig. 8. DBSP/P200 spectra and their corresponding cutout images of the two dual quasars. The slit placement is labelled as yellow lines.

Fig. 9. DBSP/P200 spectra and their corresponding cutout images of the two projected quasars and one star-star system (J0621+5227). The slit placement is labelled as yellow lines.

Fig. 10. Blue side of the spectrum of J1929+6009B spans from 3500 Å to 5250 Å. The solid orange line represents the absorptions at z = 2.797. From left to right, the features are Lyβ, OVI, Lyα, and NV doublet.

Acknowledgments

This research uses data obtained through the Telescope Access Program (TAP), which has been funded by the TAP association, including Centre for Astronomical Mega-Science CAS(CAMS), XMU, PKU, THU, USTC, NJU, YNU, and SYSU. We thank the anonymous referee for very valuable and constructive comments that help us to improve the paper significantly. We thank Jiao Li, Chao Liu for insightful discussions. We thank astropy, pandas, matplotlib, astroscrappy, lenstronomy for providing convenient and reliable python packages. Z.H. acknowledges support from the China Postdoctoral Science Foundation under Grant Number GZC20232990 and the National Natural Science Foundation of China (Grant No. 12403104). R.L. acknowledges the support of the National Nature Science Foundation of China (No 12203050). Y.S. acknowledges the support from the National Science Foundation of China (12333001) and the China Manned Space Program through its Space Application System. G.L. acknowledges the support of the China Manned Spaced Project (CMS-CSST-2021-A12). N.L. acknowledges the support of the CAS Project for Young Scientists in Basic Research (No. YSBR-062), the science research grants from the China Manned Space Project (No. CMS-CSST-2021-A01), and the hospitality of the International Centre of Supernovae (ICESUN), Yunnan Key Laboratory at Yunnan Observatories, Chinese Academy of Sciences. D.D.S. acknowledges the support from the National Science Foundation of China (12303015) and the National Science Foundation of Jiangsu Province (BK20231106). This project used data obtained with the Dark Energy Camera (DECam), which was constructed by the DES collaboration. Funding for the DES Projects has been provided by the U.S. Department of Energy, the U.S. National Science Foundation, the Ministry of Science and Education of Spain, the Science and Technology Facilities Council of the United Kingdom, the Higher Education Funding Council for England, the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign, the Kavli Institute of Cosmological Physics at the University of Chicago, Center for Cosmology and Astro-Particle Physics at the Ohio State University, the Mitchell Institute for Fundamental Physics and Astronomy at Texas A&M University, Financiadora de Estudos e Projetos, Fundacao Carlos Chagas Filho de Amparo, Financiadora de Estudos e Projetos, Fundacao Carlos Chagas Filho de Amparo a Pesquisa do Estado do Rio de Janeiro, Conselho Nacional de Desenvolvimento Cientifico e Tecnologico and the Ministerio da Ciencia, Tecnologia e Inovacao, the Deutsche Forschungsgemeinschaft and the Collaborating Institutions in the Dark Energy Survey. The Collaborating Institutions are Argonne National Laboratory, the University of California at Santa Cruz, the University of Cambridge, Centro de Investigaciones Energeticas, Medioambientales y Tecnologicas-Madrid, the University of Chicago, University College London, the DES-Brazil Consortium, the University of Edinburgh, the Eidgenossische Technische Hochschule (ETH) Zurich, Fermi National Accelerator Laboratory, the University of Illinois at Urbana-Champaign, the Institut de Ciencies de l’Espai (IEEC/CSIC), the Institut de Fisica d’Altes Energies, Lawrence Berkeley National Laboratory, the Ludwig Maximilians Universitat Munchen and the associated Excellence Cluster Universe, the University of Michigan, NSF’s NOIRLab, the University of Nottingham, the Ohio State University, the University of Pennsylvania, the University of Portsmouth, SLAC National Accelerator Laboratory, Stanford University, the University of Sussex, and Texas A&M University.

References

Appendix A: Light and mass modelling

We use the two-Moffat profile I_P, i (i presents the i-th image) to describe the light distributions of the quasars in each band, which can be considered as PSFs in DESI-LS images. I_P, i is characterized by the following parameters: the width parameters (α₁, α₂), shape parameters (β₁, β₂), position angles (ϕ_m, 1, ϕ_m, 2), offsets between the centres of the first and second Moffat profiles (δx_m, δy_m), the amplitude ratio between the first and second Moffat profiles (f), the amplitude of the first Moffat profile (A_mⁱ), and the coordinates of the centre of the first Moffat profile (x_mⁱ, y_mⁱ). We determine the parameters α₁, α₂, β₁, β₂, ϕ_m, 1, ϕ_m, 2, δx_m, δy_m, and f through star fitting using a nearby star close to the system of interest, and leaving A_mⁱ, x_mⁱ, and y_mⁱ to be fitted in later procedures. We employ a single Sérsic profile along with i PSFs to model the light distribution of lensed quasar systems. The model can be expressed as:

$$\begin{array}{c}\hfill I_{\mathit{model}}=I_s+{\mathrm{\Sigma }}_{i=1}^{\mathit{nimg}}{I}_{P,i}.\end{array}$$(A.1)

The Sérsic profile (I_s) is characterized by seven parameters: Sérsic index (n_s), half-light radius (r_s), position angle (ϕ_s), axis-ratio (q_s), amplitude (A_s) and the coordinates of the lens centre (x_lens, y_lens). Initially, we fit the band exhibiting the most prominent lensing galaxy signal (as indicated by the residual image of Legacy Survey Imaging) to minimise A_mⁱ, x_mⁱ, and y_mⁱ of the PSF model and x_lens, y_lens of Sersic model use the following likelihood function.

$$\begin{array}{c}\hfill \text{log}p(I_{\mathit{data}}|I_{\mathit{model}})=\frac{{(I_{\mathit{data}}-I_{\mathit{model}})}^2}{2{\sigma }_{\mathit{bkg}}^2},\end{array}$$(A.2)

where σ_bkg denotes the background noise. Subsequently, we fit the other bands using the same procedure. In those fittings, we fix x_mⁱ, y_mⁱ, x_lens, and y_lens according to the results obtained from the first band. The other parameters (A_mⁱ, n_s, r_s, ϕ_s, q_s, and A_s) were allowed to be differed among each bands. We adopt a SIE as our lens mass model since it has been shown to align well with observations across numerous studies (see e.g. Bolton et al. 2012; Sonnenfeld et al. 2013). During the lens modelling process, we assume that the SIE halo shares the same centre as the Sérsic profile previously determined (x_lens, y_lens). Here, to avoid the effect of microlensing and AGN light-variation, we only fit the positions of multiple images, which are established during the light modelling phase. The parameters adjusted in mass modelling include: Einstein radius (θ_E), axis ratio and position angle of SIE (q_SIE, ϕ_SIE), and source position (x_src, y_src). We note that q_SIE and ϕ_SIE were constrained within a range centered on q_s and ϕ_s, which were determined during the light modeling stage based on the results from the band with the highest signal-to-noise ratio for the lensing galaxy. To be specific, q_SIE was allowed for a ±30% variation compared to q_s, while ϕ_SIE was allowed for a ±30° variation compared to ϕ_s. Totally, we used four inputted parameters (x_m¹, x_m², y_m¹, y_m²) to constrain five parameters (θ_E, x_src, y_src, q_SIE, and ϕ_SIE), and gave q_SIE and ϕ_SIE priors. The likelihood function used in SIE fitting is:

$$\begin{array}{c}\hfill \text{log}p({\mathit{r}}_i|{\mathit{r}}_i^P)={\displaystyle \sum _i}\frac{{|{\mathit{r}}_i-{\mathit{r}}_i^P|}^2}{{\sigma }_{p,i}^2},\end{array}$$(A.3)

with

$$\begin{array}{c}\hfill {\mathit{r}}_i=\left(\begin{array}{c}{x}_m^i\\ y_m^i\end{array}\right),\end{array}$$(A.4)

where σ_p, i represents the astrometric error of i-th image, ${\mathit{r}}_i^P$ is the position of multiple images predicted by SIE model. After obtaining all parameters, we can further estimate the time delay by varying the range of z_d, set between 0.1 and 1.25 to encompass the z_d of the majority known lensed quasars. The time delay is defined as:

$$\begin{array}{c}\hfill t_d=arrival\_time({\mathit{Img}}_B)-arrival\_time({\mathit{Img}}_A),\end{array}$$(A.5)

where labels A and B correspond to those in Figure 4. In Figure 5, we illustrate the time delay versus different values of z_d.

Fig. A.1. Data for four dual quasars, verified using publicly available spectral datasets. On the left, we provide essential details: Δv represents the velocity difference, ‘sep’ denotes the image separation of the two quasars, and ‘dis’ indicates the projected separation in the line-of-sight direction at lower redshift. The second column displays the spectra, where the fluxes have been smoothed using a Gaussian kernel with a standard deviation of 5 Å. The third column features the grz colour image from the DESI-LS DR10, while the final column shows the residual image derived from DESI-LS DR10.

Fig. A.2. Data for nine projected quasars, verified using publicly available spectral datasets. The configuration of this plot is the same with Figure A.1.

Fig. A.2. continued.

All Tables

All Figures

	Fig. 1. DESI-LS colour images of the 10 targets observed by DBSP/P200 during Oct 15–16, 2023. The cutout size of each image is 12″ × 12″ and north is up, while east is left. These cutout images are obtained from Legacy Survey Imaging DR10 in grz bands and re-plotted in this work using HumVI (Marshall et al. 2016).
In the text

	Fig. 2. DBSP/P200 spectra and their corresponding cutout images of the two confirmed lensed quasars. The slit placement is labelled using the yellow lines.
In the text

	Fig. 3. Image decomposition results in grz-band for our confirmed lensed quasars. First: J0746+1344. Second: J2121−0826. First column (A): Original data. Second (B): Light model for two images. Third (C): A-B; Fourth (D): Light model for best-fitted lensing galaxy assuming Sérsic profile. Fifth (E): A-B-D.
In the text

	Fig. 4. Modelled caustic and critical curves, along with the positions of images for the two confirmed lenses shown in Figure 2, are compared with the observed positions of multiple images. The plots are centred around the lens light centre, which is determined through light modelling and provided in Table 4.
In the text

	Fig. 5. Time delay as a function of zd of two confirmed lensed quasars.
In the text

	Fig. 6. DBSP/P200 spectra and their corresponding cutout images of the three likely lensed quasars. The slit placement is labelled using the yellow lines.
In the text

Fig. 7. Colour-colour plot for J0122−1344 is displayed, with blue points indicating stars. The yellow and purple pentagrams represent J0122−1344A based on Legacy Survey Imaging photometry. Furthermore, the yellow and purple triangles correspond to J0122−1344A and J0122−1344B, respectively, as determined by the light modelling results from this study, which utilises two PSFs, along with one Sérsic profile (refer to Appendix A).

In the text

	Fig. 8. DBSP/P200 spectra and their corresponding cutout images of the two dual quasars. The slit placement is labelled as yellow lines.
In the text

	Fig. 9. DBSP/P200 spectra and their corresponding cutout images of the two projected quasars and one star-star system (J0621+5227). The slit placement is labelled as yellow lines.
In the text

	Fig. 10. Blue side of the spectrum of J1929+6009B spans from 3500 Å to 5250 Å. The solid orange line represents the absorptions at z = 2.797. From left to right, the features are Lyβ, OVI, Lyα, and NV doublet.
In the text

Fig. A.1. Data for four dual quasars, verified using publicly available spectral datasets. On the left, we provide essential details: Δv represents the velocity difference, ‘sep’ denotes the image separation of the two quasars, and ‘dis’ indicates the projected separation in the line-of-sight direction at lower redshift. The second column displays the spectra, where the fluxes have been smoothed using a Gaussian kernel with a standard deviation of 5 Å. The third column features the grz colour image from the DESI-LS DR10, while the final column shows the residual image derived from DESI-LS DR10.

In the text

	Fig. A.2. Data for nine projected quasars, verified using publicly available spectral datasets. The configuration of this plot is the same with Figure A.1.
In the text

	Fig. A.2. continued.
In the text