Abstract

We have developed a wideband receiver system for simultaneous observations in CO lines of J = 2–1 and J = 3–2 transitions using the Osaka 1.85 m mm–submm telescope. As a frequency separation system, we developed multiplexers that connect three types of diplexers, each consisting of branch-line couplers and high-pass filters. The radio frequency (RF) signal is eventually distributed into four frequency bands, each of which is fed to a superconductor–insulator–superconductor (SIS) mixer. The RF signal from the horn is divided into two frequency bands by a wideband diplexer with a fractional bandwidth of |$56\%$|⁠, and then each frequency band is further divided into two bands by each diplexer. The developed multiplexers were designed, fabricated, and characterized using a vector network analyzer. The measurement results showed good agreement with the simulation. The receiver noise temperature was measured by connecting the SIS-mixers, one of which has a wideband 4–21 GHz intermediate frequency (IF) output. The receiver noise temperatures were measured to be ∼70 K in the 220 GHz band, ∼100 K in the 230 GHz band, 110–175 K in the 330 GHz band, and 150–250 K in the 345 GHz band. This receiver system has been installed on the 1.85 m telescope at the Nobeyama Radio Observatory. We succeeded in simultaneous observations of six CO isotopologue lines with the transitions of J = 2–1 and J = 3–2 toward the Orion KL as well as on-the-fly mappings toward the Orion KL and W 51.

1 Introduction

The 1.85 m mm–submm telescope (figure 1) of Osaka Prefecture University (OPU) has been operated at the Nobeyama Radio Observatory (NRO) to observe molecular clouds in nearby star-forming regions and along the Galactic plane in the molecular lines of 12CO, 13CO, and C18O (J = 2–1) (Onishi et al. 2013; Nishimura et al. 2015, 2020). A receiver system based on superconductor–insulator–superconductor (SIS) mixers has been installed in the telescope in order to achieve highly sensitive and efficient observations with low noise temperature close to the several times quantum limit. Now, we are planning to relocate the 1.85 m telescope to a site at the Atacama Desert in Chile (altitude ∼2500 m) and to install a new wideband receiver for simultaneous observations of CO isotopologue lines with the transitions of J = 2–1, J = 3–2, and other molecular lines. Historically, the quasi-optical system has been used to achieve simultaneous observations in such widely separated molecular lines in frequency (e.g., Kramer et al. 1998). This is because the large fractional bandwidth and wideband intermediate frequency (IF) system had been difficult to develop with only the waveguide system. As a frequency separation system, several types of multiplexers have been proposed and developed so far (Han et al. 2013; Okada et al. 2020; Kojima et al. 2017; Nakajima et al. 2020). Han et al. (2013) and Okada et al. (2020) have developed receivers using quasi-optical diplexers. The quasi-optical diplexer enables simultaneous observations of molecular spectra with widely separated frequencies. However, the optics becomes large and complicated, which makes the system difficult to implement inside the front-end cryostat or even into the telescope optical path. Therefore, quasi-optical diplexers are usually installed at room temperature, with a consequent increase in noise temperature due to extra loss in the insertion path. Kojima et al. (2017) and Nakajima et al. (2020) have developed waveguide multiplexers composed of several diplexers. These multiplexers made the receiver system compact and simple since the system can be realized with a single horn. Therefore, we decided to develop a waveguide multiplexer that separates 230 GHz and 345 GHz bands, a wideband optical system, and a corrugated horn for the 210–375 GHz band. The development of the wideband optics with a corrugated horn is described in Yamasaki et al. (2020, 2021).

Osaka 1.85 m mm–submm telescope with radome (a) and without radome (b). (Color online)
Fig. 1.

Osaka 1.85 m mm–submm telescope with radome (a) and without radome (b). (Color online)

In recent years, the National Astronomical Observatory of Japan (NAOJ) developed a wideband SIS-mixer with a 4–21 GHz IF output (Kojima et al. 2020). This wideband IF system will enhance the flexibility of the IF system and thus the ability to observe many scientifically desired molecular lines simultaneously. For example, the 12CO, 13CO, and C18O molecular lines in 345 GHz band can be observed simultaneously in one side-band of the SIS-mixer with the 4–21 GHz IF output. Another advantage of the wide IF mixer is that the local oscillator (LO) only needs to cover a narrower frequency range for the same radio frequency (RF) coverage. We decided to use this SIS-mixer to achieve our aimed observations.

In this paper, we report the development of the wideband receiver system using the waveguide multiplexers, including the wideband SIS-mixer with an IF output of 4–21 GHz, and the measurement results of the system. In addition, we also report the results of the first light observations with the 1.85 m telescope using the new wideband receiver system.

2 Development of a wideband waveguide multiplexer system

2.1 Introduction to the waveguide multiplexer system

There are several types of waveguide multiplexers, such as hybrid-coupled, manifold-coupled, circulator-coupled, and directional filter multiplexers. Cameron and Yu (2007) described a comparison of these multiplexers. Manifold-coupled, circulator-coupled, and directional filter types (Ruiz-Cruz et al. 2008; Zhang et al. 2018) are compact and have several advantages over the hybrid-coupled type. However, these multiplexers have a complex structure, narrow bandwidth, and high insertion loss.

The hybrid-coupled multiplexer can be made by cascade connections of diplexers (Asayama et al. 2015; Hasegawa et al. 2017) that consist of a branch-line coupler (BLC) and a band-pass filter (BPF). These multiplexers were designed to separate a fractional bandwidth of about |$30\%$| (between minimum frequency and maximum frequency) in the RF band into three or four bands (Kojima et al. 2017; Nakajima et al. 2020) to achieve simultaneous observations of molecular lines. In their multiplexers, the signals are extracted in the order of high to low frequency. This means that lower-frequency signals pass through more diplexers, indicating that the lower the frequency band, the longer the path length and thus the greater the insertion loss. Since the waveguide insertion loss is lower at lower frequencies, this compensates to some extent the longer path length, obtaining a flatter insertion loss within the RF band.

2.2 Conceptual design of the wideband waveguide multiplexer

Here, we show the conceptual design of the waveguide multiplexer with a fractional bandwidth of more than |$50\%$|⁠. As shown in figure 2, the wideband diplexer splits the RF signal from the horn into two frequency bands with a wide bandwidth. Then, each frequency band is divided into two bands by each diplexer. This makes it possible to unify the number of diplexers to be passed through to two for all channels, in contrast to the cascade connection type. The multiplexers consist of the wideband diplexer (α) to separate the two RF bands, 210–275 GHz and 275–375 GHz, and the following two diplexer (β) and diplexer (γ) focused on lines of CO isotopologues with the transitions of J = 2–1 and J = 3–2, respectively, to achieve dual side-band observations for each RF band. The wideband SIS-mixers with 4–21 GHz IF output (Kojima et al. 2020) are expected to be connected to four output channels of the wideband multiplexer and driven by two LO sources at fixed frequencies of 240 GHz and 325 GHz.

Schematic diagram of the final configuration of the planned new receiver system for the simultaneous observations in CO isotopologue lines of the transitions of J = 2–1 and J = 3–2. The receiver is mainly composed of three components; a wideband corrugated horn, a waveguide multiplexer, and wideband IF SIS-mixers. The waveguide multiplexer consists of a wideband diplexer (α) and two diplexers (β, γ). The wideband diplexer (α) provides frequency separation between 210–275 GHz and 275–375 GHz. The two output ports are connected to each diplexer (β) and (γ), respectively, and eventually separated into four frequency bands: 215–235, 245–265, 305–321, and 329–345 GHz. (Color online)
Fig. 2.

Schematic diagram of the final configuration of the planned new receiver system for the simultaneous observations in CO isotopologue lines of the transitions of J = 2–1 and J = 3–2. The receiver is mainly composed of three components; a wideband corrugated horn, a waveguide multiplexer, and wideband IF SIS-mixers. The waveguide multiplexer consists of a wideband diplexer (α) and two diplexers (β, γ). The wideband diplexer (α) provides frequency separation between 210–275 GHz and 275–375 GHz. The two output ports are connected to each diplexer (β) and (γ), respectively, and eventually separated into four frequency bands: 215–235, 245–265, 305–321, and 329–345 GHz. (Color online)

It has been quite challenging to develop a wideband diplexer (α) that can separate the fractional bandwidth over |$50\%$|⁠. In Gonzalez et al. (2017), a prototype wideband diplexer with combining LO signal was developed and fabricated in the 275–500 GHz band. Referring to the model of Gonzalez et al. (2017), we optimized our system as an RF circuit of 210–375 GHz band. The design of the wideband diplexer, including the wideband BLC and high-pass filter (HPF), is described in detail in Masui et al. (2019).

Each output port of diplexer (β) and diplexer (γ) is connected to the IF 4–21 GHz SIS-mixer. Therefore, it is necessary to design the diplexers with a passband of 21 GHz. In addition, the passbands of the diplexers are set to be larger than 21 GHz in order to accommodate further expansion of the IF band of the SIS-mixer. To achieve this condition, we applied a HPF-type diplexer (Hasegawa et al. 2017), not a BPF-type diplexer used in conventional multiplexers.

2.3 Design and specifications of the waveguide multiplexer

Figure 3 shows the schematic diagram of an HPF–type diplexer, consisting of two BLCs and two HPFs. The BLCs divide the input radio waves into the same intensity of −3 dB with a phase difference of 90°, while the HPFs pass the radio waves higher than the cutoff frequency and reflect the lower ones. Therefore, when radio waves are input to the diplexer from port 1, the radio waves higher than the cutoff frequency is output from port 4, and the radio waves lower than the cutoff frequency is output from port 2. Port 3 is terminated by a 4-K load that absorbs unwanted RF signal and LO power. The load contributes to the receiver noise temperature on ports 2 and 4 as a source of RF thermal noise. Since the temperature is cooled to 4 K, the load has little effect on the noise temperature; the thermal noise is very weak compared to the intensity of the incoming radio signal from port 1. In order to design three diplexers that make up the wideband multiplexer, the BLC and HPF of each are simulated and optimized.

Schematic diagram of the diplexer we developed. It is composed of two BLCs and two high-pass filters (HPFs). (Color online)
Fig. 3.

Schematic diagram of the diplexer we developed. It is composed of two BLCs and two high-pass filters (HPFs). (Color online)

First, the design of the BLC is explained. Here, the characteristic is described in scattering parameters (S-parameters: Gonzalez 1997). As mentioned above, in a BLC, it is important to divide the input intensity in half, but it is not possible to make both |$S_{31}$| and |$S_{41}$| precisely equal to −3.0 dB at all frequencies. In this development, we designed the wideband diplexer (α) BLC with |$S_{31}$| and |$S_{41}$| to be −3.0 ± 0.75 dB, and the diplexer (β) and (γ) BLC with |$S_{31}$| and |$S_{41}$| to be −3.0 ± 0.5 dB. Furthermore, the return losses |$S_{11}$| and |$S_{21}$| of each BLC were set to be better than 20 dB. The target characteristics of each BLC are summarized in table 1. We optimized the design of the diplexers by analyzing the S-parameters using the electromagnetic field analysis software HFSS. Figure 4 shows the simulation results of these BLCs. In the optimized design of the wideband BLC (α), |$S_{31}$| and |$S_{41}$| are simulated to be −3.0 ± 0.81 dB, and |$S_{11}$| and |$S_{21}$| are better than ∼21 dB; for BLC (β) and (γ), |$S_{31}$| and |$S_{41}$| are −3.0 ± 0.54 dB and −3.0 ± 0.41 dB, respectively, and |$S_{11}$| and |$S_{21}$| are better than ∼22 dB.

(Top) Conceptual structure of a BLC. The number and width of branches vary from circuit to circuit. (α) The simulation results of the wideband BLC, (β) 210–275 GHz band BLC, and (γ) 290–375 GHz band BLC. The sizes of the BLCs are shown in detail in the Appendix. (Color online)
Fig. 4.

(Top) Conceptual structure of a BLC. The number and width of branches vary from circuit to circuit. (α) The simulation results of the wideband BLC, (β) 210–275 GHz band BLC, and (γ) 290–375 GHz band BLC. The sizes of the BLCs are shown in detail in the Appendix. (Color online)

Table 1.

Goal characteristics of the BLCs.

ComponentParameterGoal
Wideband BLC (α)Δ = S31S411.5 dB
S11, S21∼20 dB
200 GHz BLC (β)Δ = S31S411.0 dB
S11, S21∼20 dB
300 GHz BLC (γ)Δ = S31S411.0 dB
S11, S21∼20 dB
ComponentParameterGoal
Wideband BLC (α)Δ = S31S411.5 dB
S11, S21∼20 dB
200 GHz BLC (β)Δ = S31S411.0 dB
S11, S21∼20 dB
300 GHz BLC (γ)Δ = S31S411.0 dB
S11, S21∼20 dB
Table 1.

Goal characteristics of the BLCs.

ComponentParameterGoal
Wideband BLC (α)Δ = S31S411.5 dB
S11, S21∼20 dB
200 GHz BLC (β)Δ = S31S411.0 dB
S11, S21∼20 dB
300 GHz BLC (γ)Δ = S31S411.0 dB
S11, S21∼20 dB
ComponentParameterGoal
Wideband BLC (α)Δ = S31S411.5 dB
S11, S21∼20 dB
200 GHz BLC (β)Δ = S31S411.0 dB
S11, S21∼20 dB
300 GHz BLC (γ)Δ = S31S411.0 dB
S11, S21∼20 dB

Secondly, we discuss the design of the three types of HPFs; (α), (β), and (γ). To design the wideband HPF (α), a compact cavity filter (Rosenberg et al. 2004) was applied to achieve a large passband, sharper cutoff, and low insertion loss. The cutoff frequency was set to ∼275 GHz, and the return loss was set to be better than 20 dB. The cutoff frequencies of HPFs (β) and (γ) were set to ∼238 GHz and ∼323.5 GHz, respectively, and the return loss was set to be better than 20 dB. The target characteristics of each HPF are summarized in table 2. The models of the HPFs (α), (β), and (γ) are shown in figure 5, and the simulation results are presented below. In the final design of the wideband HPF (α), the cutoff frequency is about 277 GHz, and the return loss is better than ∼21 dB; for HPFs (β) and (γ), the return losses are better than ∼20 dB in most of the bands. In part of the frequency range of HPF (β), the return loss was ∼18 dB.

Model and simulation results of the designed HPFs. (α) the HPF of the wideband diplexer (α), (β) the 210–275 GHz band HPF, and (γ) the 290–375 GHz HPF. The sizes of the HPFs are shown in detail in the Appendix. (Color online)
Fig. 5.

Model and simulation results of the designed HPFs. (α) the HPF of the wideband diplexer (α), (β) the 210–275 GHz band HPF, and (γ) the 290–375 GHz HPF. The sizes of the HPFs are shown in detail in the Appendix. (Color online)

Table 2.

Goal characteristics of the HPFs.

ComponentParameterGoal
Wideband HPF (α)Return loss∼20 dB
Cutoff frequency∼275 GHz
200 GHz HPF (β)Return loss∼20 dB
Cutoff frequency∼238 GHz
300 GHz HPF (γ)Return loss∼20 dB
Cutoff frequency∼323.5 GHz
ComponentParameterGoal
Wideband HPF (α)Return loss∼20 dB
Cutoff frequency∼275 GHz
200 GHz HPF (β)Return loss∼20 dB
Cutoff frequency∼238 GHz
300 GHz HPF (γ)Return loss∼20 dB
Cutoff frequency∼323.5 GHz
Table 2.

Goal characteristics of the HPFs.

ComponentParameterGoal
Wideband HPF (α)Return loss∼20 dB
Cutoff frequency∼275 GHz
200 GHz HPF (β)Return loss∼20 dB
Cutoff frequency∼238 GHz
300 GHz HPF (γ)Return loss∼20 dB
Cutoff frequency∼323.5 GHz
ComponentParameterGoal
Wideband HPF (α)Return loss∼20 dB
Cutoff frequency∼275 GHz
200 GHz HPF (β)Return loss∼20 dB
Cutoff frequency∼238 GHz
300 GHz HPF (γ)Return loss∼20 dB
Cutoff frequency∼323.5 GHz

The configuration of the multiplexer was made to connect each designed BLC to the HPF as shown in figure 6. In order to connect the developed multiplexers to the SIS-mixers and to mix them with LO in the SIS-mixers, LO couplers were built into the input ports of the diplexers (β) and (γ). LO couplers have a coupling ratio of about 17 dB and a directivity of more than 20 dB in the frequency range of interest. The input LO signals are transmitted to each output port of the diplexers (β) and (γ) in the same way as RF signals. The size of the input / output waveguide is set to WR-3.15 (0.8 mm × 0.4 mm) in order to prevent higher-order modes of the waveguide from occurring; the cutoff frequencies of the fundamental and the first higher-order modes are 187.5 GHz and 375 GHz, respectively. Note that the “WR” designation stands for “rectangular waveguides”, and the number that follows “WR” is the width of the waveguide opening in millimetres, divided by 10 (Hesler et al. 2007). To connect these components, impedance-matching circuits are connected. The simulation results are shown in figure 7, where the four frequency bands corresponding to IF 4–21 GHz are shown by the blue areas. This analysis is calculated using the conductivity of a perfect electrical conductor (PEC). The total return loss is simulated to be ∼15 dB. We also simulated the image rejection ratios, and they are designed to be better than ∼20 dB in most of the bands. The sizes of the diplexers are shown in detail in the Appendix.

Layout of the multiplexer. It consists of a wideband 210–375 GHz diplexer (α), a 210–275 GHz diplexer (β), and a 290–375 GHz diplexer (γ). This shows how the three diplexers are connected. Enlarged views of the diplexers are also shown. A waveguide circuit that supplies LO is connected to each of the two diplexers (β) and (γ), which is then mixed with the RF signal and fed to the connected SIS-mixers. (Color online)
Fig. 6.

Layout of the multiplexer. It consists of a wideband 210–375 GHz diplexer (α), a 210–275 GHz diplexer (β), and a 290–375 GHz diplexer (γ). This shows how the three diplexers are connected. Enlarged views of the diplexers are also shown. A waveguide circuit that supplies LO is connected to each of the two diplexers (β) and (γ), which is then mixed with the RF signal and fed to the connected SIS-mixers. (Color online)

Simulation results of the designed multiplexer. The conductivity of the PEC is assumed in the simulation. The blue square area indicates the bandwidths to be used in the observations. (Color online)
Fig. 7.

Simulation results of the designed multiplexer. The conductivity of the PEC is assumed in the simulation. The blue square area indicates the bandwidths to be used in the observations. (Color online)

2.4 Fabrication of waveguide multiplexer

The wideband diplexer (α) and the diplexers (β) and (γ) were fabricated in separate blocks to accommodate future changes, such as the observed frequency. Each block is composed of two E-plane split-blocks, and the waveguide circuits were fabricated by direct machining using an end mill. Aluminum alloy A6063, which is easy to machine and has excellent electrical conductivity, was used for these split-blocks. In particular, the conductivity of A6063 at 4 K is ∼4.0 × 108 S m−1 (Hatch 1984), which reduces the insertion loss of the waveguide circuit significantly at low temperatures. An anti-cocking type waveguide flange was used. The size of the wideband diplexer (α) and diplexer (β) blocks was 20 mm × 20 mm × 20 mm when the two split-blocks were assembled. The size of the diplexer (γ) block was set to 20 mm × 22 mm × 20 mm when assembled. Figure 8a)shows a photograph of the manufactured diplexer (α). Figure 8c)shows an optical image focused on the circuit of the diplexer (α) taken with a Zygo NewView 8000, and figure 8b)shows the waveguide multiplexer with these diplexers assembled.

(a) Photograph of the split-blocks, described as Block A and Block B, of the fabricated wideband diplexer (α). (b) Photograph of the assembled wideband multiplexer taken alongside a 100 Japanese Yen coin. (c) Detailed optical image of the circuit part. (Color online)
Fig. 8.

(a) Photograph of the split-blocks, described as Block A and Block B, of the fabricated wideband diplexer (α). (b) Photograph of the assembled wideband multiplexer taken alongside a 100 Japanese Yen coin. (c) Detailed optical image of the circuit part. (Color online)

2.5 Measurement system of waveguide multiplexer

A PNA-X vector network analyzer (VNA) and waveguide extenders, VNA extension modules, from the National Institute of Information and Communications Technology (NICT) were used for the multiplexer and waveguide transitions measurements. To ensure a wide frequency range of 210–375 GHz, a WR-3.4 (0.864 mm × 0.432 mm) extender was used for 220–330 GHz, and a WR-2.2 (0.559 mm × 0.279 mm) extender was used for 325–500 GHz. In order to minimize the mismatch at the flange interface with the extenders, tapered transitions with waveguide sizes from WR-3.15 to WR-3.4 and from WR-3.15 to WR-2.2 with a length of 20 mm were prepared. The relatively long waveguide was chosen for this measurement because of its good return loss characteristics. A picture of a typical measurement setup for back-to-back measurements of the S-parameters of the WR-3.15 to WR-2.2 transition using the WR-2.2 extender is shown in figure 9a. Prior to the measurements, a short-open-load-through (SOLT) calibration—a standard calibration method of the VNA and extenders—of two ports was performed. The measurement results of the waveguide transitions with each extender are shown in figure 9b: the return loss and insertion loss for the WR-3.15 to WR-3.4 transition are 28 dB and 2 dB, respectively, and those for the WR-3.15 to WR-2.2 transition are better than 20 dB and 2.5 dB, respectively.

(a) Photograph of the setup to measure S-parameters for the WR-3.15 to WR-2.2 waveguide transition in the back-to-back configuration using the WR-2.2 VNA extender. (b) Measurement results of the WR-3.15 to WR-3.4 transition and the WR-3.15 to WR-2.2 transition using WR-3.4 and WR-2.2 extenders, respectively. (Color online)
Fig. 9.

(a) Photograph of the setup to measure S-parameters for the WR-3.15 to WR-2.2 waveguide transition in the back-to-back configuration using the WR-2.2 VNA extender. (b) Measurement results of the WR-3.15 to WR-3.4 transition and the WR-3.15 to WR-2.2 transition using WR-3.4 and WR-2.2 extenders, respectively. (Color online)

2.6 Measurement results of the waveguide multiplexer

The measurement results of the multiplexer are shown in figure 10. For S21, S31, S41, and S51, the insertion loss of the waveguide transition described in the previous section are subtracted. The simulation results using the conductivity of PEC are also plotted as solid lines. The cutoff frequency of the diplexer (β) in the 210–275 GHz band is shifted to a higher frequency of about 3 GHz. We performed the 3-D measurement of the waveguide and found that the waveguide size became smaller, which caused the frequency shift of 3 GHz. The return loss was measured to be ∼15 dB, which is consistent with the simulation results.

Comparison of the simulation and the measurement of the S-parameters of the multiplexer. The measurement was carried out using VNA, and the results are indicated by open circles. The simulation assumes the conductivity of PEC, and the results are shown by solid lines. (a) The output levels to Ports 2, 3, 4, and 5 are expressed as S21, S31, S41, and S51, respectively. (b) The return loss, S11. The simulation (solid line) and measurement results (open circles) show good agreement, and the return loss is measured to be better than about 15 dB throughout the required frequency bands. (Color online)
Fig. 10.

Comparison of the simulation and the measurement of the S-parameters of the multiplexer. The measurement was carried out using VNA, and the results are indicated by open circles. The simulation assumes the conductivity of PEC, and the results are shown by solid lines. (a) The output levels to Ports 2, 3, 4, and 5 are expressed as S21, S31, S41, and S51, respectively. (b) The return loss, S11. The simulation (solid line) and measurement results (open circles) show good agreement, and the return loss is measured to be better than about 15 dB throughout the required frequency bands. (Color online)

Here, we discuss the insertion loss by also taking the surface roughness into account. We assume the surface roughness of the waveguide to be 0.075 μm from Kojima et al. (2017) and use an aluminum conductivity of 3.8 × 107 S m−1. The aluminum A6063 is known to exhibit good conductivity at low temperatures of 4 K (Hatch 1984). With the above-assumed surface roughness, we simulated the insertion loss using the HFSS both at room temperature and 4 K, and plot the results in figure 11 with the measurement results. This figure shows that the simulation using room temperature reproduces the measurement results well, indicating that the surface roughness of the waveguide is estimated to be ∼0.075 μm. At 4 K with the surface roughness, the simulation results show that the insertion loss of the multiplexer is estimated to be ∼1 dB. For example, if the SIS-mixer is used with an SSB noise temperature of about 60 K, the effect of 1 dB can be calculated to be ∼15 K, indicating that the effect of insertion loss of the multiplexer is small, though not negligible.

Detailed view of the measured S21, S31, S41, and S51 with a simulation assuming a room temperature/4 K conductivity and the realistic roughness of the waveguide surface. The open circles are the measurement results. The solid lines are the results from the simulation assuming the 300-K conductivity and the 0.075 μm surface roughness. The measurement results shown with open circles, measured at 300 K, are found to delineate the solid lines well. The dotted lines are the simulation results with the conductivity at 4 K of aluminum and the roughness of the waveguide surface 0.075 μm. This indicates that the S-parameters of the manufactured multiplexer at 4 K are about 1 dB over the necessary frequency band. (Color online)
Fig. 11.

Detailed view of the measured S21, S31, S41, and S51 with a simulation assuming a room temperature/4 K conductivity and the realistic roughness of the waveguide surface. The open circles are the measurement results. The solid lines are the results from the simulation assuming the 300-K conductivity and the 0.075 μm surface roughness. The measurement results shown with open circles, measured at 300 K, are found to delineate the solid lines well. The dotted lines are the simulation results with the conductivity at 4 K of aluminum and the roughness of the waveguide surface 0.075 μm. This indicates that the S-parameters of the manufactured multiplexer at 4 K are about 1 dB over the necessary frequency band. (Color online)

3 Application to the 1.85 m telescope: simultaneous observations of CO in 230 GHz and 345 GHz bands

3.1 Receiver configuration

Due to the limited availability of some of the components, we modified some parts of the final configuration of figure 2 as shown below. Using the developed multiplexer and the SIS-mixer developed by Kojima et al. (2020), we developed a wideband receiver for simultaneous observations of CO isotopologue lines in the 230 GHz and 345 GHz bands. Because the RF band of the developed IF 4–21 GHz SIS-mixer supports the frequency range of 275–500 GHz, it is used in the 345 GHz band of our receiver. In the 230 GHz band, we used an SIS-mixer with IF 4–8 GHz which employs a parallel-connected twin junction with a mirror symmetrical circuit about the bisection plane in the center of the waveguide (Asayama et al. 2003). However, the multiplexer developed here is designed to use wideband IF even in the 230 GHz band, so it cannot be used with a side-band-separated SIS-mixer in the 4–8 GHz band. Therefore, instead of the diplexer (β), we decided to use another diplexer developed by Hasegawa et al. (2017), and change the LO frequency in the 230 GHz band to 225 GHz. The corrugated horn, which couples the incoming 210–375 GHz RF signal from the telescope to the waveguide input for the multiplexer is described in detail in another paper (Yamasaki et al. 2021). A photograph of the RF circuit for the simultaneous CO observations is shown in figure 12, and the schematic diagram of the receiver for the observations is shown in figure 13. One of the output ports of the diplexer (γ) in the 300–325 GHz frequency band was terminated because there is no desired CO line.

Photograph of the receiver system assembled. The system consists of a wideband corrugated horn, the wideband multiplexer we developed here, a wideband SIS-mixer with 4–21 GHz IF output, and a conventional SIS-mixer with 4–8 GHz IF output.
Fig. 12.

Photograph of the receiver system assembled. The system consists of a wideband corrugated horn, the wideband multiplexer we developed here, a wideband SIS-mixer with 4–21 GHz IF output, and a conventional SIS-mixer with 4–8 GHz IF output.

Schematic diagram of the wideband receiver system for the observations. (Color online)
Fig. 13.

Schematic diagram of the wideband receiver system for the observations. (Color online)

The LO system is a Phase Matrix FSW 0020, and the active multiplier in the 230 GHz band is an hxi HAFME 4-139 × 4 multiplier and VDI WR-4.3 output tripler, while in the 345 GHz band a Quinstar QMM-933516060 × 6 multiplier and VDI WR-3.4 output tripler are used. In our system, a BPF is inserted between the signal generator and the active multiplier chain to reduce the LO noise, and the details will be presented in another paper. The LO frequencies were set to 225 GHz and 324.9 GHz for the frequency bands of the diplexers (β) and (γ), respectively. The 230 GHz SIS-mixers’ outputs are amplified at 4 K by 4–8 GHz cryogenic low noise amplifiers (CLNA), manufactured by Nitsuki, with cryogenic isolators, and feed into the first IF chains as shown in figure 13 at room temperature. The IF output is then downconverted by the second mixers with the appropriate bandpass filters before each mixer, and the downconverted signals in the 0–2 GHz range are transferred through the second IF chains, as shown in figure 13, to the spectrometers. The 345 GHz SIS-mixer’s output is amplified at 4 K by 3–22 GHz CLNAs manufactured by Low Noise Factory. A 10-dB attenuator is inserted after the first CLNA instead of the wideband isolator covering the 3–22 GHz which is not currently available. The attenuator is able to avoid the saturation of the 2nd CLNA due to the wideband signal, and we install it at 4 K to minimize the excess noise. Then, the signal feeds into the first IF chain at room temperature. A coaxial diplexer is attached to the room-temperature part of the wideband IF 4–21 GHz, and only the required bandwidth is extracted by BPFs. The extracted signals are downconverted to the 0–2 GHz range by each mixer and also transferred through the second IF chains shown in figure 13 to the spectrometers. High-resolution wideband fast Fourier transform spectrometers (XFFTS; Klein et al. 2012) were used to obtain the spectra in the frequency range of 0–2 GHz in the second IF chain. The resulting inputs are as follows: board 1, 13CO and C18O (J = 2–1); board 2, 12CO (J = 2–1); board 3, 13CO and C18O (J = 3–2); and board 4, 12CO (J = 3–2).

3.2 Laboratory measurement of the receiver noise temperatures with 4–21 GHz IF output SIS-mixer

Figure 13 shows a schematic diagram of the wideband receiver. In the laboratory measurement, the single side-band (SSB) noise temperature was measured using a spectrum analyzer in the first IF section in order to check the characteristics of the entire IF as well as at the actual observed frequency. The receiver noise temperature was measured using the standard Y–factor method with room temperature (“Hot”; THot = 290 K) and liquid nitrogen cooled (“Cold”; TCold = 77 K) blackbody loads. Each blackbody radiation was input using an elliptical mirror for laboratory measurements; a Keysight N9938A spectrum analyzer was used to check the IF output. Figure 14 shows the results of the receiver noise temperature measurements in the laboratory. The x-axis corresponds to the first IF frequency, the left-hand side of the y-axis shows the calculated noise temperature value, and the right-hand side of the y-axis shows the IF output power when Hot and Cold signals are input. The noise temperature of the entire IF is about 70 to 100 K in the 220 GHz band, about 100 K in the 230 GHz band, and about 100 to 200 K in the 330 to 345 GHz band. The blue area corresponds to the spectrometer frequencies. The noise temperatures of the double side-band (DSB) of individual SIS-mixers were also measured separately before the assembly. The double of the DSB temperature, the SSB equivalent noise temperature, is shown by the black dotted line in figure 14. As estimated in subsection 2.6, we expect an extra noise corresponding to ∼1 dB due to the insertion loss of the multiplexer. However, the SSB equivalent noise temperature found using the DSB measurement is quite similar to that of the multiplexer system. One of the possible causes is that the SIS-mixers’ performance tuning might be different between the measurements. An independent measurement of the same SIS-mixers at NAOJ with the different measurement system resulted in the best DSB noise temperature of ∼45 K within the same frequency range at the LO frequency of 325 GHz, which is slightly better than the present DSB measurement (cf. figure 14c). The optical misalignment at the present DSB measurement might also be a possible cause of the higher noise temperature. Although the present measurements are not sensitive enough to estimate the insertion loss of the multiplexer, the overall performance of the system is reasonably high and seems to be consistent with the simulation result, at least not causing the significant increase of the noise temperatures of the system.

Measurement results of the SSB receiver noise temperature. The horizontal axis shows the first IF frequency. Each graph shows the measurement results for (a) a lower side-band (LSB) with the LO frequency of 225 GHz, (b) an upper side-band (USB) with the same LO frequency, and (c) an USB with the LO frequency 324.9 GHz. The orange lines show IF output power when observing a Hot load, the blue lines IF output power when observing a Cold load, and the green line shows receiver noise temperature calculated from the Hot and Cold measurement. The grey dotted lines show the SSB equivalent noise temperatures found by doubling the DSB temperature of each SIS-mixer measured separately. The blue areas correspond to the spectrometer frequencies. (Color online)
Fig. 14.

Measurement results of the SSB receiver noise temperature. The horizontal axis shows the first IF frequency. Each graph shows the measurement results for (a) a lower side-band (LSB) with the LO frequency of 225 GHz, (b) an upper side-band (USB) with the same LO frequency, and (c) an USB with the LO frequency 324.9 GHz. The orange lines show IF output power when observing a Hot load, the blue lines IF output power when observing a Cold load, and the green line shows receiver noise temperature calculated from the Hot and Cold measurement. The grey dotted lines show the SSB equivalent noise temperatures found by doubling the DSB temperature of each SIS-mixer measured separately. The blue areas correspond to the spectrometer frequencies. (Color online)

4 Commissioning observations using the developed receiver system on the 1.85 m telescope

The wideband receiver system is installed on the Osaka 1.85 m mm–submm telescope at NRO, which is developed and operated by OPU. We also developed the antenna/receiver control system using distributed processing, which will allow us to respond flexibly to the introduction of new instruments. The details of telescope control and antenna drive system can be found in Kondo et al. (2020). We observed molecular clouds in Orion KL (⁠|$\mbox{RA}= {05^{\rm h}35^{\rm m}14{_{.}^{\rm s}}5}$|⁠, |$\mbox{Dec}= -{05^{\circ}22^{\prime }30^{\prime \prime }}$| at J2000.0) and succeeded in simultaneous observations of 12CO, 13CO, and C18O (J = 2–1 and J = 3–2). Figure 15 shows the chopper wheel calibrated spectra of six CO emission lines. By comparison with the intensity of 12CO (J = 2–1) in the 230 GHz band given by Nishimura et al. (2015), the main beam efficiency is considered to be comparable to the previous one. The beam size and main beam efficiency are described in detail in Yamasaki et al. (2021). The on-the-fly (OTF) mapping result toward the Orion KL region is shown in figure 16; although C18O (J = 3–2) was too weak to be mapped, we succeeded in the simultaneous OTF observations of the other five emission lines. The OTF mapping result toward the W 51 region is also shown in figure 17.

Observation results toward Orion KL. These CO lines with the transitions of J = 2–1 and J = 3–2 were observed simultaneously. The x-axis is the velocity of local standard of rest, and the y-axis is the antenna temperature. Spurious lines appear around 50 km s−1 of C18O (J = 2–1) and −20 km s−1 of 13CO (J = 3–2). (Color online)
Fig. 15.

Observation results toward Orion KL. These CO lines with the transitions of J = 2–1 and J = 3–2 were observed simultaneously. The x-axis is the velocity of local standard of rest, and the y-axis is the antenna temperature. Spurious lines appear around 50 km s−1 of C18O (J = 2–1) and −20 km s−1 of 13CO (J = 3–2). (Color online)

OTF maps of the peak brightness temperature toward Orion KL. The map size is 1° × 1°. White circles show the beam size. Note that the C18O (J = 3–2) lines are too faint to be mapped. (Color online)
Fig. 16.

OTF maps of the peak brightness temperature toward Orion KL. The map size is 1° × 1°. White circles show the beam size. Note that the C18O (J = 3–2) lines are too faint to be mapped. (Color online)

OTF maps of the peak brightness temperature toward W 51. The map size is 1° × 1°. White circles show the beam size. (Color online)
Fig. 17.

OTF maps of the peak brightness temperature toward W 51. The map size is 1° × 1°. White circles show the beam size. (Color online)

5 Conclusion

We have developed a waveguide-based wideband multiplexer covering a frequency range of 210–375 GHz for the simultaneous observations of CO isotopologue molecular lines of J = 2–1 and 3–2 transitions. The results of this study can be summarized as follows:

  1. We have designed and developed a wideband diplexer (α) with a fractional bandwidth of |$56\%$|⁠. The return loss is better than ∼17 dB, and the cutoff frequency for dividing the RF signal is 275 GHz. The basic design is a combined BLC and HPF diplexer proposed by Hasegawa et al. (2017), but to achieve a wider passband, sharper cutoff, and lower insertion loss, a compact cavity filter (Rosenberg et al. 2004) was used.

  2. The diplexers (β) and (γ) were developed to further divide the radio waves into two parts (210–275 GHz and 275–375 GHz bands, respectively). They are also a combination of BLC and HPF and are designed to have a return loss of better than 20 dB.

  3. The three diplexers listed above were combined to form a multiplexer. The multiplexer can divide radio waves in the frequency band 210–375 GHz into four bands: 215–235 GHz [corresponding to C18O, 13CO and 12CO (J = 2–1)], 245–265 GHz, 305–321 GHz, and 329–345 GHz [C18O, 13CO, and 12CO (J = 3–2)]. We measured the S-parameters of the multiplexer using a vector network analyzer and found that the return loss was better than 15 dB in all frequency bands. The insertion loss was also consistent with the simulation results and was estimated to be approximately 1 dB at 4 K.

  4. The waveguide circuit was designed so that the output from the diplexers (β) and (γ) and the LO signal are mixed and then input to the SIS-mixer. The output port from the diplexer (γ) in the 329–345 GHz band is connected to an SIS-mixer (Kojima et al. 2017) with an IF output bandwidth of 4–21 GHz. This enables simultaneous detection of CO 3 emission lines of the J = 3–2 transition. The noise temperature of the entire IF is measured to be about 70 to 100 K in the 220 GHz band, about 100 K in the 230 GHz band, and 100 to 200 K in the 330 to 345 GHz band.

  5. Using the receiver systems on the 1.85 m radio telescope, we succeeded in simultaneous observations of six CO isotopologue molecular (J = 2–1, 3–2) lines toward Orion KL. This is also the first astronomical observation result with the SIS-mixer with an IF output bandwidth of 4–21 GHz. The OTF mapping of 1° × 1° around Orion KL was also successfully done in five emission lines except for C18O (J = 3–2), which is the weakest and noisiest emission line due to the atmosphere. In addition, the OTF mapping toward W 51 with the same mapping size was successfully done in six emission lines.

In the ALMA Development Roadmap,1 which describes a roadmap for future developments of ALMA, the wideband RF and IF are listed as one of the important development items. Our study pursues wideband heterodyne reception in mm and submm wavelengths, and actual observation using a telescope has been achieved. We believe that this research result is a fundamental achievement for future wideband observations by high-performance radio telescopes such as ALMA.

Funding

This work was supported by JSPS KAKENHI Grant Numbers JP15K05025, JP26247026, JP18H05440, and JP20J23670. This work was also supported the grant of Joint Development Research supported by the Research Coordination Committee, National Astronomical Observatory of Japan (NAOJ), National Institutes of Natural Sciences (NINS).

Acknowledgments

The authors would like to thank Issei Watanabe and Satoshi Ochiai of NICT. They helped us to measure the S-parameter of waveguide circuits. The authors would like to also thank Shigeru Fuji of NRO and Toshihisa Tsutsumi of Yamaguchi University for the kind help at Nobeyama. The authors are very grateful to the entire staff of the NRO and everyone related to the 1.85m telescope. We are grateful that Shohei Ezaki of NAOJ and Masayuki Ishino of KMCO helped us to develop the wideband diplexer (α). We thank the referee, Dr. Noriyuki Kawaguchi, for helpful comments.

Appendix. The final design of diplexers

In this appendix, the structural parameters of the developed wideband diplexer (α), (β), and (γ) are shown in table 3 and in figures 18, 19, and 20. The design of the wideband diplexer (α) uses different BLCs before and after the HPF. The designs of the above BLCs is described in detail in Masui et al. (2019).

Dimensions of the wideband diplexer (α) when the two split-blocks were assembled. (a) The top view. (b) The cross-sectional view from the side. The parameters starting with “R” stand for the corner radius. This design uses different BLCs before (left) and after (right) the HPF. The design of the above BLCs are described in detail in Masui et al. (2019).
Fig. 18.

Dimensions of the wideband diplexer (α) when the two split-blocks were assembled. (a) The top view. (b) The cross-sectional view from the side. The parameters starting with “R” stand for the corner radius. This design uses different BLCs before (left) and after (right) the HPF. The design of the above BLCs are described in detail in Masui et al. (2019).

Dimensions of the diplexer (β) when the two split-blocks were assembled. (a) The top view. (b) The cross-sectional view from the side. The parameters starting with “R” stand for the corner radius. The diplexer (β) was designed to be symmetric.
Fig. 19.

Dimensions of the diplexer (β) when the two split-blocks were assembled. (a) The top view. (b) The cross-sectional view from the side. The parameters starting with “R” stand for the corner radius. The diplexer (β) was designed to be symmetric.

Dimensions of the diplexer (γ) when the two split-blocks were assembled. (a) The top view. (b) The cross-sectional view from the side. The parameters starting with “R” stand for the corner radius. The diplexer (γ) was designed to be symmetric.
Fig. 20.

Dimensions of the diplexer (γ) when the two split-blocks were assembled. (a) The top view. (b) The cross-sectional view from the side. The parameters starting with “R” stand for the corner radius. The diplexer (γ) was designed to be symmetric.

Table 3.

Size parameters of the diplexers.

ParameterFigure 18 [mm]Figure 19 [mm]Figure 20 [mm]
x11.0921.0920.8
x20.740.800.065
x30.600.730.55
x40.500.630.466
x50.80NANA
x60.74NANA
y12.732.732.246
y20.400.180.34
y30.320.280.34
y40.052.52.0
y50.26NANA
y60.14NANA
y72.33NANA
z10.300.300.4
z20.280.280.34
z30.220.250.29
z40.160.150.18
z50.58NANA
h10.250.300.20
h20.20NANA
w10.050.0750.116
w20.063NANA
L10.190.300.2
L20.20NANA
L30.22NANA
L40.22NANA
R10.0250.050.05
R20.05NANA
ParameterFigure 18 [mm]Figure 19 [mm]Figure 20 [mm]
x11.0921.0920.8
x20.740.800.065
x30.600.730.55
x40.500.630.466
x50.80NANA
x60.74NANA
y12.732.732.246
y20.400.180.34
y30.320.280.34
y40.052.52.0
y50.26NANA
y60.14NANA
y72.33NANA
z10.300.300.4
z20.280.280.34
z30.220.250.29
z40.160.150.18
z50.58NANA
h10.250.300.20
h20.20NANA
w10.050.0750.116
w20.063NANA
L10.190.300.2
L20.20NANA
L30.22NANA
L40.22NANA
R10.0250.050.05
R20.05NANA
Table 3.

Size parameters of the diplexers.

ParameterFigure 18 [mm]Figure 19 [mm]Figure 20 [mm]
x11.0921.0920.8
x20.740.800.065
x30.600.730.55
x40.500.630.466
x50.80NANA
x60.74NANA
y12.732.732.246
y20.400.180.34
y30.320.280.34
y40.052.52.0
y50.26NANA
y60.14NANA
y72.33NANA
z10.300.300.4
z20.280.280.34
z30.220.250.29
z40.160.150.18
z50.58NANA
h10.250.300.20
h20.20NANA
w10.050.0750.116
w20.063NANA
L10.190.300.2
L20.20NANA
L30.22NANA
L40.22NANA
R10.0250.050.05
R20.05NANA
ParameterFigure 18 [mm]Figure 19 [mm]Figure 20 [mm]
x11.0921.0920.8
x20.740.800.065
x30.600.730.55
x40.500.630.466
x50.80NANA
x60.74NANA
y12.732.732.246
y20.400.180.34
y30.320.280.34
y40.052.52.0
y50.26NANA
y60.14NANA
y72.33NANA
z10.300.300.4
z20.280.280.34
z30.220.250.29
z40.160.150.18
z50.58NANA
h10.250.300.20
h20.20NANA
w10.050.0750.116
w20.063NANA
L10.190.300.2
L20.20NANA
L30.22NANA
L40.22NANA
R10.0250.050.05
R20.05NANA

Footnotes

1

Carpenter, J., Iono, D., Testi, L., Whyborn, N., Wootten, A., & Evans, N. 2018, ALMA Memo Series, 612.

References

Asayama
 
S.
,
Noguchi
 
T.
,
Ogawa
 
H.
 
2003
,
Int. J. Infrared Millimeter Waves
,
24
,
1091

Asayama
 
S.
,
Hasegawa
 
Y.
,
Mizuno
 
A.
,
Ogawa
 
H.
,
Onishi
 
T.
 
2015
,
Int. J. Infrared Millimeter Waves
,
36
,
445

Cameron
 
R. I.
,
Yu
 
M.
 
2007
,
IEEE Microwave Magazine
,
8
,
46

Gonzalez
 
A.
,
Kojima
 
T.
,
Kaneko
 
K.
,
Asayama
 
S.
 
2017
,
IEEE Trans. Terahertz Sci. Technol.
,
7
,
669

Gonzalez
 
G.
 
1997
,
Microwave Transistor Amplifiers Analysis and Design
, 2nd ed. (
Hoboken
:
Prentice-Hall
)

Han
 
S.-T.
 et al.  
2013
,
PASP
,
125
,
539

Hasegawa
 
Y.
,
Asayama
 
S.
,
Harada
 
R.
,
Tokuda
 
K.
,
Kimura
 
K.
,
Ogawa
 
H.
,
Onishi
 
T.
 
2017
,
PASJ
,
69
,
91

Hatch
 
J. E.
ed.
1984
,
Aluminum: Properties and Physical Metallurgy
(
Materials Park
:
Aluminum Association Inc. and ASM International
)

Hesler
 
J. L.
,
Kerr
 
A. R.
,
Grammer
 
W.
,
Wollack
 
E.
 
2007
, in
Proc. 30th Int. Symp. on Space THz Technol
, ed. A. Karpov (
Red Hook
:
Curran Associates, Inc.
),
100

Klein
 
B.
,
Hochgürtel
 
S.
,
Krämer
 
I.
,
Bell
 
A.
,
Meyer
 
K.
,
Güsten
 
R.
 
2012
,
A&A
,
542
,
L3

Kojima
 
T.
,
Gonzalez
 
A.
,
Asayama
 
S.
,
Uzawa
 
Y.
 
2017
,
IEEE Trans. Terahertz Sci. Technol.
,
7
,
10

Kojima
 
T.
,
Kiuchi
 
H.
,
Uemizu
 
K.
,
Uzawa
 
Y.
,
Kroug
 
M.
,
Gonzalez
 
A.
,
Dippon
 
T.
,
Kageura
 
T.
 
2020
,
A&A
,
640
,
L9

Kondo
 
H.
 et al.  
2020
,
Proc. SPIE
,
11452
,
114523M

Kramer
 et al.  
1998
,
Proc. SPIE
,
3357
,
711

Masui
 
S.
 et al.  
2019
, in
Proc. 30th Int. Symp. on Space THz Technol
, ed. A. Karpov (
Red Hook
:
Curran Associates, Inc.
),
73

Masui
 
S.
 et al.  
2020
,
Proc. SPIE
,
11453
,
114534F

Nakajima
 
T.
,
Haratani
 
K.
,
Mizuno
 
A.
,
Suzuki
 
K.
,
Kojima
 
T.
,
Uzawa
 
Y.
,
Asayama
 
S.
,
Watanabe
 
I.
 
2020
,
Int. J. Infrared Millimeter Waves
,
41
,
1530

Nishimura
 
A.
 et al.  
2015
,
ApJS
,
216
,
18

Nishimura
 
A.
 et al.  
2020
,
Proc. SPIE
,
11445
,
114457F

Okada
 
N.
 et al.  
2020
,
PASJ
,
72
,
7

Onishi
 
T.
 et al.  
2013
,
PASJ
,
65
,
78

Rosenberg
 
U.
,
Amari
 
S.
,
Bornemann
 
J.
,
Vahldieck
 
R.
 
2004
, in
Proc. 34th European Microwave Conf.
, Vol.
2
(
New York
:
IEEE
),
985

Ruiz-Cruz
 
J. A.
,
Montejo–Garai
 
J. R.
,
Rebollar
 
J. M.
,
Sobrino
 
S.
 
2008
,
Progr. Electromag. Res.
,
86
,
39

Yamasaki
 
Y.
 et al.  
2020
,
Proc. SPIE
,
11453
,
114534H

Yamasaki
 
Y.
 et al.  
2021
,
PASJ
,
73
,
1116

Zhang
 
Y.-B.
,
Chen
 
B.
,
Ran
 
C.-Z.
 
2018
,
IEICE Electronics Exp.
,
15
,
20180826

This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://dbpia.nl.go.kr/journals/pages/open_access/funder_policies/chorus/standard_publication_model)