N3AR: Dual-Band mm-VLBI Receiver

• N3AR is a 3 mm receiver operating over 67–116 GHz that enables dual-frequency observations through a shared-optical-path design.
• Its innovative dichroic beamsplitter setup allows simultaneous low- and high-frequency data collection, crucial for frequency phase transfer in VLBI.
• Commissioning results show dramatic improvements in coherence time and baseline sensitivity, paving the way for high-resolution mm-VLBI imaging.

Searching arXiv for N3AR and related VLBI/FPT work to ground the article in the cited literature. N3AR is a 3 mm receiver at the APEX telescope, operating over 67–116 GHz and commissioned in October 2024. Its defining role, as described in “Shared-optical-path VLBI frequency phase transfer from 86 to 258 GHz on an 8600 km baseline: Demonstrated with the APEX and IRAM 30 m telescopes” (Zhao et al., 14 Jul 2025), is to add a new low-frequency band for APEX while enabling simultaneous dual-frequency observations through a shared-optical-path configuration with a dichroic beamsplitter. In that configuration, N3AR provides the low-frequency channel for frequency phase transfer (FPT) from 86 to 258 GHz, supports APEX participation in the existing 3 mm global VLBI network, and serves as a technical platform for extending high-frequency VLBI coherence, sensitivity, and imaging fidelity.

1. Instrument identity and core specifications

N3AR operates in the 67–116 GHz RF band, with an IF band of 1.5–20 GHz delivered as four 18-GHz-wide sub-bands (Zhao et al., 14 Jul 2025). The receiver is a cooled HEMT front-end in a single Dewar, and it uses a sideband-separating room-temperature downconverter (RPG module) that yields two sidebands per polarization. Its polarization basis is natively dual linear, while a retractable quarter-wave plate made of grooved Teflon converts the signal to dual circular for VLBI. The reported typical single-sideband receiver noise temperature, $T_{\rm nrec}$ (DSB), is 40–60 K across most of the band, measured by the hot/cold-load method.

A concise summary of the reported hardware parameters is given below.

Parameter	Reported value	Function
RF band	67–116 GHz	3 mm observing band
IF band	1.5–20 GHz	Output delivered as four 18-GHz-wide sub-bands
Polarization	Native dual linear; dual circular for VLBI via quarter-wave plate	Supports VLBI recording modes
Typical $T_{\rm nrec}$ (DSB)	40–60 K	Receiver noise across most of the band

The calibration system uses hot (348 K) and ambient cold (290 K) loads that are switched optically. The local oscillator is an 8–20 GHz synthesizer with an $\times 8$ multiplier, and continuous phase-cal, described as “tone” injection, is provided through a Schottky-diode comb. These specifications place N3AR at the intersection of heterodyne instrumentation and VLBI backend calibration, rather than as a standalone continuum receiver.

2. Optical layout and shared-optical-path implementation

The distinctive architectural feature of N3AR is its integration into a shared-optical-path (SOP) system (Zhao et al., 14 Jul 2025). A single tertiary relay directs the beam into the Cassegrain cabin with 12.5 dB edge taper and $4\omega$ clearance. Within that common optical train, a wire-grid dichroic composed of two cross-wired grids with 500 $\mu$m wire pitch and optimized grid separation reflects 67–116 GHz to N3AR and transmits 196–345 GHz to nFLASH-230 or SEPIA-345.

The dichroic insertion is fully within the common optical train, so both bands track identical atmospheric paths. This is the key optical precondition for FPT: atmospheric phase perturbations measured at the lower frequency are relevant to the higher-frequency signal because the two receivers sample the same line of sight through the troposphere. The simplified path is reported as

$${\rm M1} \rightarrow {\rm M2} \rightarrow {\rm M3}\;({\rm tertiary\ mirror}) \rightarrow {\rm dichroic},$$

with reflection to N3AR and transmission to the higher-frequency receiver in Nasmyth A.

Approximate beam-splitting coefficients are also reported. The reflection coefficient is $R_{\rm pol} \simeq 0.98$ at 100 GHz, and the transmission is $T \simeq 0.95$ at 230 GHz. The detailed design is stated to follow classical wire-grid theory. In practical terms, this optical arrangement makes APEX a simultaneous low-/high-frequency station rather than a telescope that alternates between bands.

3. System performance formalism

The commissioning overview gives the system-temperature and baseline-sensitivity relations explicitly (Zhao et al., 14 Jul 2025). The system temperature is parameterized in terms of receiver temperature, spillover and forward efficiencies, atmospheric opacity, elevation, atmospheric emission, and the cosmic microwave background. The baseline RMS noise is written as

$$\Delta S \;=\; \frac{2\,k}{A_{\rm eff}} \;\frac{T_{\rm sys}}{\sqrt{2\,B\,t_{\rm int}}}\,,$$

where $k$ is Boltzmann’s constant, $T_{\rm nrec}$0 is antenna effective area, $T_{\rm nrec}$1 is the recording bandwidth, and $T_{\rm nrec}$2 is the coherent integration time.

Within this formalism, coherent integration time is a primary control variable for sensitivity. Because the reported observing mode uses $T_{\rm nrec}$3 GHz, any increase in coherence time achieved through FPT translates directly into a lower $T_{\rm nrec}$4 on a single baseline. This connection underlies the subsequent emphasis on phase stabilization rather than solely on receiver-noise reduction.

4. Dual-frequency VLBI configuration and FPT procedure

The demonstration experiment used APEX, at 5104 m in Chile, equipped with N3AR and nFLASH230, together with the IRAM 30 m telescope at Pico Veleta in Spain using EMIR, on a baseline of approximately 8,623 km (Zhao et al., 14 Jul 2025). The reference frequencies were $T_{\rm nrec}$5 GHz and $T_{\rm nrec}$6 GHz, with the integer ratio

$T_{\rm nrec}$7

Data were recorded with 1.024 GHz bandwidth in dual polarization, circular at APEX and linear at IRAM, and correlated with DiFX at 0.5 MHz spectral and 0.4 s temporal resolution. Phase calibration differed by station: APEX injected a continuous comb into both bands, whereas IRAM EMIR used loaded-tone calibration between scans. Fringe fitting in AIPS used SOLINT=10 s with overlap=5 s, and multiband delays and rates were solved independently at each band.

The FPT procedure is summarized in three steps. First, one solves the low-frequency phase $T_{\rm nrec}$8. Second, one scales by $T_{\rm nrec}$9 and subtracts from the observed high-frequency phase $\times 8$0:

$\times 8$1

Third, one applies the residual delays and rates from the low band to the high band. The method therefore treats the 86 GHz signal as a simultaneously measured atmospheric phase reference for 258 GHz.

5. Commissioning results and coherence behavior

The principal commissioning result is that the phase fluctuations at 86 and 258 GHz were found to be well correlated, and that applying FPT caused the interferometric phases at the higher frequency to vary much more slowly (Zhao et al., 14 Jul 2025). The reported example is a 5.5 min scan on CTA 102 at 23:07:30 UT, with elevation $\times 8$2 at APEX and $\times 8$3 at IRAM.

Before FPT, the 258 GHz phase $\times 8$4 varied by $\times 8$5 rad min$\times 8$6, corresponding to a range of approximately 26 rad per scan. The coherence time, defined operationally by the point at which the coherence factor falls below 0.9, was $\times 8$7 s. After FPT, the corrected phase $\times 8$8 varied by only $\times 8$9 rad min$4\omega$0, with a range of approximately 1.2 rad per scan, and the coherence was extended to $4\omega$1 s.

The coherence factor is defined as

$4\omega$2

The tabulated values for the CTA 102 scan are as follows.

$4\omega$3 (s)	$4\omega$4	$4\omega$5
10	0.88	0.99
20	0.61	0.97
60	0.32	0.92
300	0.05	0.81

Similar improvements were reported for 3C 454.3 and PKS B0420-014, and faint targets became fringe-detectable only after FPT. The most direct quantitative implication drawn in the source is that extending coherent integration from 10 s to 60 s yields a $4\omega$6 baseline-sensitivity improvement at 258 GHz.

6. Scientific and technical implications

The reported sensitivity gain is only one part of the claimed impact (Zhao et al., 14 Jul 2025). The same analysis states that longer coherent time and more baselines equipped with SOP receivers would produce dynamic-range gains exceeding an order of magnitude, which is identified as crucial for EHT and GMVA imaging of black-hole shadows and jet cores. In astrometric terms, residual dispersive terms such as ionospheric effects and instrumental drift are described as varying on timescales of minutes or longer; combining FPT with source-frequency phase referencing (SFPR) is therefore expected to deliver $4\omega$7as registration between bands.

The future hardware path is explicit. The same shared-optics concept is projected to enable simultaneous 86/345 GHz operation with N3AR and nFLASH345, while multi-site SOP arrays involving KVN, the ALMA Phased Array, LMT, NOEMA, and APEX are identified as a route toward pushing FPT to 430 GHz and beyond. The science areas named in that context are AGN core shifts, black-hole photon rings, and ultra-compact maser astrometry.

These developments position N3AR less as an isolated receiver upgrade than as an enabling component in a broader mm/sub-mm VLBI calibration strategy. A plausible implication is that its long-term importance will be determined not only by its intrinsic receiver performance, but by how extensively shared-optical-path dual-frequency operation is adopted across the global network.

7. Place within APEX and mm-VLBI infrastructure

N3AR adds a low-frequency 3 mm band to APEX and allows the telescope to join the existing 3 mm global VLBI network (Zhao et al., 14 Jul 2025). In the reported commissioning mode, it functions jointly with a higher-frequency receiver through the dichroic beamsplitter, so that APEX can operate as a simultaneous dual-band VLBI station rather than as a single-band participant. The commissioning summary states that N3AR has been fully commissioned in a shared-optics configuration and that the first intercontinental 86 to 258 GHz FPT demonstration shows dramatic coherence and sensitivity enhancements.

From an instrumentation perspective, the decisive feature is the combination of a broadband 3 mm HEMT receiver, SOP optics, dual-polarization capability adaptable for VLBI, and continuous phase-cal injection. From a network perspective, the receiver provides the low-frequency anchor required for atmospheric phase transfer to substantially higher frequencies. In that sense, N3AR occupies a specific infrastructural niche: it is the APEX 3 mm frontend through which simultaneous multi-band VLBI calibration becomes operationally practical.