1. Introduction

The energy budget, dynamics and chemical cycle of the Venusian atmosphere is strongly influenced by the H2SO4-H2O clouds, which float at around 45–70 km altitudes (Esposito et al., 1983). H2SO4 is thought to be produced photochemically near the cloud top via the oxidation of SO2, which is abundant below the cloud top, and thus the clouds basically have the characteristics of photochemical aerosols. On the other hand, a strong coupling between cloud condensation and atmospheric motion is expected to occur in the lower part of the cloud layer, where the heating of clouds by upwelling infrared radiation drives vertical convection (Baker and Schubert, 1992; Imamura and Hashimoto, 2001).

The cloud layer absorbs solar radiation, thereby driving various atmospheric motions that might play a central role in the momentum balance of the super-rotation. The super-rotation is a planet-wide easterly wind on Venus; the wind speed increases with height at all latitudes and reaches about 100 m s−1 near the cloud top (Schubert et al., 1980). Various mechanisms explaining the super-rotation have been proposed to date (Gierasch et al., 1997). Among them is the combination of a thermally-driven Hadley circulation and large-scale eddies which transports angular momentum equatorward, causing a net upward transport of angular momentum (Gierasch, 1975; Rossow and Williams, 1979; Iga and Matsuda, 2005). Another candidate is acceleration by thermal tides, which are excited in the cloud layer by periodic solar heating and which propagate vertically to induce momentum exchange between atmospheric layers (Fels and Lindzen, 1974; Newman and Leovy, 1992; Takagi and Matsuda, 2005, 2007). Diurnal and semidiurnal tides are observed in the temperature structure above clouds (Taylor et al., 1980; Zasova et al., 2002; Tellmann et al., 2009) and in the cloud-top wind field (Rossow et al., 1990). The dissipation, at cloud heights, of Kelvin waves that originate in the lower atmosphere might also accelerate the atmosphere (Del Genio and Rossow, 1990; Yamamoto and Tanaka, 1997). Characterization of the dynamics at cloud heights is crucial for understanding the super-rotation.

The cloud-top wind field has been observed by tracking the movements of small-scale ultraviolet features seen in the dayside cloud images taken from Venus orbiters (e.g., Rossow et al., 1990; Moissl et al., 2009). These observations revealed a mean poleward flow with a velocity of up to 10 m s−1 at the cloud top. However, the limitation of the local time coverage may allow a significant contamination of the thermal tide component in the estimated meridional circulation, which can be as large as 10 m s−1 (Newman and Leovy, 1992). Characterization of the meridional circulation requires wind measurements at all local times.

Eddy motions at the cloud top have also been studied, based mostly on the observations of ultraviolet contrasts, which show various morphological features whose origins are mostly unknown. Among these are the cell-like structures seen in the sub-solar region (Rossow et al., 1980; Markiewicz et al., 2007); their horizontal scales of up to several hundreds of kilometers are too large for convective cells, and the cloud-top region is considered to be stably stratified (Baker and Schubert, 1992; Toigo et al., 1994). This mysterious feature might be related to the vertical transport of cloud materials and might ultimately determine the cloud structure.

Fig. 1
figure 1

The flight model of LIR before being assembled to the spacecraft. The sensor unit (a), and the power supply unit (b).

It should be noted that these ultraviolet features do not necessarily represent the cloud structure, but reflect the horizontal inhomogeneities of ultraviolet absorbers. The solar ultraviolet radiation scattered by the cloud top shows absorption by SO2 at wavelengths shorter than 320 nm and by unknown materials at longer wavelengths (Esposito et al., 1997). The mixing ratios of both SO2 and the unknown absorber are considered to increase precipitously with decreasing altitude below the cloud top (Pollack et al., 1980; Bertaux et al., 1996), and thus the spatial distributions of these species should also be sensitive to vertical air motions. Such vertical winds may also influence the cloud-top height, but the relation of the cloud height to the vertical wind is uncertain. Recently, cloud altimetry using scattered near-infrared solar radiation showed that ultraviolet dark features tend to correspond to higher clouds in the southern high latitude (Ignatiev et al., 2009). Methods to map the cloud-top height over broad local times and latitudinal regions would further constrain the cloud dynamics.

Efforts to map the thermal emission from the cloud top of Venus have been made using ground-based telescopes (e.g., Apt et al., 1980) and infrared radiometers onboard the Pioneer Venus orbiter (Taylor et al., 1980) and Venera 15 and 16 (Zasova et al., 2007). They revealed the structures of thermal tides and other planetary-scale waves as well as the polar dipole structure. The visible and Infrared Thermal Imaging Spectrometer (VIRTIS) onboard Venus Express revealed that the temperature distribution in the south polar region is similar to that in the north polar region (Piccioni et al., 2007). However, the observations have been limited in spatial resolution, temporal resolution and the latitudinal coverage, preventing the studies of mesoscale processes and the derivation of wind field by cloud tracking.

Akatsuki, which is the first Japanese Venus orbiter, aims at understanding the atmospheric dynamics and cloud physics of Venus. Akatsuki maps clouds and minor constituents successively with four cameras covering wavelengths from infrared to ultraviolet, detects lightning flashes with a high-speed photometer, and retrieves the vertical structure of the atmosphere with a radio occultation technique (Nakamura et al., 2007, 2011). The planned orbit around Venus is a 30-hour-period elliptical orbit (370–78,500 km) near the ecliptic plane, and cloud images will be obtained every 2 hours for each observation wavelength. Science instruments altogether observe multiple height levels of the atmosphere to model the three-dimensional structure and dynamics. Among them is the Longwave Infrared Camera (LIR), which maps the thermal emission from the cloud top at 8–12 µm wavelengths (Taguchi et al., 2007). Unlike other cameras onboard Akatsuki, LIR is able to take images of dayside and nightside clouds with equal quality. This is advantageous not only to the studies of the diurnal cycles of cloud processes but also to the precise determination of the zonal-mean meridional circulation.

2. Instrument Design

The global-averaged temperature field obtained from the Venera 15 IR spectrometry data (Zasova et al., 2007) suggests that the temperature range of the structure is between 205 and 260 K, and the numerical study of penetrative convection about the Venusian cloud between 40 and 60 km altitude has derived potential temperature profiles which suggest the range of cloud-top altitude is within several kilometers (Baker et al., 1998; Ignatiev et al., 2009). In order to detect a cloud-height difference of a few hundred meters, LIR requires 0.3 K of noise equivalent temperature difference (NETD), which can be regarded as the temperature resolution, for a 230 K target as shown in Table 1. An uncooled micro-bolometer array (UMBA) (Tanaka et al., 2000) for a commercial infrared camera is used for LIR as an image sensor. The commercial camera is designed to give the best performance when it views room-temperature objects (Wada et al., 1998). For LIR, the electronics and the driving parameters must be optimized for the low-temperature targets of this mission. The UMBA has 328×248 pixels with a pixel size of 37 µm, and its temperature is stabilized at 313 K by a Peltier temperature control system. Since an UMBA does not need cryogenic apparatus, which is commonly used for photodiode-type infrared detectors onboard spacecraft, LIR is relatively small and light in weight; it achieves a size of 200× 130× 110 mm, a weight of 3.5 kg and a power consumption of 29 W. UMBAs have been applied to a couple of space missions: the ISIR instrument onboard Space Shuttle (STS-85) (Lancaster et al., 2001) and the THEMIS instrument onboard Mars Odyssey (Christensen et al., 2004). LIR will be a benchmark of an infrared camera with an UMBA for future space explorations to Mars, asteroids and the Earth.

Table 1 Specification of LIR.

The internal structure and block diagram of LIR are shown in Figs. 2 and 3, respectively. LIR consists of a sensor unit LIR-S which manages the function of image acquisition, an external power supply unit LIR-AE which converts the primary electric power to several voltages to distribute them to a regulator and a mechanical shutter, and a baffle which keeps direct sunlight away from the optical aperture. The optics in the sensor unit consists of three germanium lenses with the F-number of 1.4. The field-of-view of 16.4° × 12.4° is based on a common requirement with other cameras. The pixel resolution of 0.05° corresponds to 26–70 km on the Venus surface when the spacecraft views Venus from distances of 3–8× 104 km on an elliptical orbit. The temperature of the optics is kept within 293–308 K by the heater-controlling electronics (HCE) of the spacecraft in order to prevent thermal distortions. The mechanical shutter which works not only as a sunlight shield but also as a blackbody for calibration is positioned just in front of the UMBA and is driven by a stepping motor. A bandpass filter whose profile is shown in Fig. 4 is inserted at the pupil position of the optics. The UMBA for LIR has large pixel-to-pixel inhomogeneities of offset and sensitivity. Although these inhomogeneities are partly reduced in the analog circuit before analog-to-digital conversion, by using the on-chip fixed pattern noise (OFPN) data which is preacquired before launch, they still remain in the raw data. In order to remove the pixel-dependent offset, LIR acquires a target data and a shutter data sequentially, and the shutter data is subtracted from the target data as shown in Fig. 5. All target and shutter images that are used in this subtraction are generated by averaging up to 128 raw images (first averaging), that are acquired continuously at a frame rate of 60 Hz. Then up to 32 offset-removed images are acquired within several minutes and averaged (second averaging). The stability of the spacecraft’s attitude, which is designed for the other cameras onboard Akatsuki, is within 0.015 deg (Nakamura et al., 2011). Since they have a higher spatial resolution than LIR, there is no restriction for the observation of LIR.

Fig. 2
figure 2

Schematic of internal structure of the sensor unit of LIR.

Fig. 3
figure 3

The block diagram of LIR when configurated in the space craft.

Fig. 4
figure 4

The transmittance spectrum for the bandpass filter incorporated in LIR. The angle of incidence is zero.

Fig. 5
figure 5

Example of infrared images acquired by the bread-board model of LIR. A raw target image (a), a raw shutter image for calibration (b), and the resultant image produced by subtracting the shutter image from the target image (c).

The control of LIR for sequential image acquisition and the subsequent data processing such as averaging and offset subtraction are performed by the Digital Electronics (DE), which is an onboard controller for four cameras including LIR (Nakamura et al., 2007). The amount of data transfer to the ground station is reduced by DE using a lossless data compression algorithm called HIREW (Takada et al., 2007). There still remains a slight inhomogeneity of the sensitivity, even in a subtracted image. This will be corrected after being transmitted to the ground. The resultant image, which gives the difference of the brightness temperature between the object and the shutter, is further converted to a brightness temperature map by adding the temperature of the shutter to all pixel values.

3. Evaluation Test of the Flight Model

3.1 Measurement of the NETD

The NETD of LIR has been evaluated in a vacuum chamber in which the pressure was maintained in the range between 10−6 and 10−8 torr. The configuration of equipment in the chamber is shown in Fig. 6. An aluminum table, whose temperature is controlled by liquid nitrogen and an electric heater, is placed at the bottom in the chamber. An aluminum plate fixed on the temperature-controlled table works as a blackbody with an emissivity of 0.85–0.92 at room temperature. The blackbody is radiatively coupled with the temperature-controlled table and shrouded by aluminum plates for stabilizing the temperature within 230±0.1 K. LIR was placed above the aluminum plate shroud, and oriented downward. The operational heat of LIR is removed to the table via aluminum joints and kept at 300 K, which is a nominal temperature in the orbit.

Fig. 6
figure 6

Vacuum chamber in which equipments are configured for image acquisition by LIR.

Two kinds of blackbodies were used for the evaluation: one is a combination of two panels which are arranged side by side and maintained at slightly different temperatures using a film heater, and the other is a plate with a homogeneous temperature. Figure 7(a) and (b) show images of these blackbodies without the secondary image accumulation. Brightening of the edges of the images is obvious both in (a) and (b). Usually in visible light photographs, vignetting of optics causes darkening of the edges of a photo. However, when a low-temperature object is imaged by infrared, the radiation emitted by the object and reaching the detector is overwhelmed by the radiation from the optics and its mount whose temperature is higher than that of the object. Therefore, vignetting results in a brightening of the image of the cold object. The temperature difference of 4.7 K between the warm (right) and cold (left) side is not apparent in (a). The vignetting effect was corrected by subtracting the image of the uniform blackbody (b). The result (Fig. 7(c)) shows that the temperature contrast was successfully detected although the boundary was blurred because of defocus. When observing Venus, the brightening of the edges will be removed by the subtraction of uniform blackbody images acquired in the vacuum chamber, or deep-space images that can be acquired during the orbit of Akatsuki.

Fig. 7
figure 7

Images of blackbodies acquired by LIR in a vacuum environment. (a) Temperature of ∼230 K with the right half being 4.7 K warmer than the left, (b) Temperature of ∼230 K without temperature contrast, (c) Same as (a), but the brightness distribution for a uniform target (b) has been subtracted.

NETD will be improved by averaging several tens of images taken within a few minutes (Geoffray et al., 2000). Figure 8 shows the NETD variations with the numbers of the first and secondary accumulations. They are estimated as NETD = dT/(S/N), where dT is the actual temperature difference between the warmer and colder blackbodies, S the difference of the observed mean brightness between the warmer and colder blackbodies in the image, and N = σ the standard deviation of the observed brightness that is evaluated locally in the warmer or colder domains. The reduction of noise by the first accumulation is effective till m = 32 as long as the number of the second accumulation n is less than 5. When the number of second averaging is five or more, the noise compression becomes negligible. This suggests that the noise is not ideally random. When the first and second averaging were 32 and 32, respectively, substituting the observed values of dT = 4.22 K, S = 2922 and N = 247 into the equation above, we obtained the best NETD ∼ 0.36 K, which meets the required specification. These numbers of averaging have been adopted for the flight model.

Fig. 8
figure 8

Variations of the NETD accompanying change of number of the first (m) and second (n) accumulation. They were estimated from images in which the temperature of the blackbodies was 230 K.

3.2 Endurance test of the shutter

The stepping motor used for driving the shutter requires to be of high reliability. Since the number of operations in a mission period is expected to be a maximum of 120,000 times, a trial endurance test has been carried out in a vacuum environment using a shutter component. The motor has attained an operation of 240,000 times for about one month with round-the-clock operation.

3.3 Tolerance of the UMBA to vibration, radiation and sunlight

The tolerance of the UMBA to the launch environment was evaluated by using a vibration testing machine. Random vibration levels that were calculated using a mathematical structure model of LIR were imposed on the UMBA for 45 seconds for three axes; the spectrally-integrated levels were several tens of Grms. The function of the UMBA was normal after the vibration test without any pixel defects or malfunctions.

Since this is the first time for this type of UMBA to be used in the space environment, tolerance of the UMBA to high-energy protons was evaluated by exposing the UMBA to a 100 MeV proton beam at the rate of 4 Gy/min with a total dose of 300 Gy which exceeds what is expected during the mission life. The UMBA was set active during the test and no malfunction was observed during the exposure to the proton beam. After the exposure, the UMBA was inspected and shown to have the nominal sensitivity, NETD and no pixel defects.

At the time of launch the shutter of LIR is closed. Vibration and shock tests show that the shutter remained closed even after the mechanical perturbations expected for the launch with margins applied. Moreover, a shutter-closing command will be issued as soon as possible after separation of the spacecraft from the launch vehicle in the case of an accidental opening of the shutter. However, there is a slight possibility that vibration or shock will cause the shutter to open at the time of launch and direct sunlight may enter the aperture of LIR for less than two minutes until attitude and orbit control is started. In such a case assumed, the tolerance of the UMBA to direct sunlight input was evaluated by using a trial piece beforehand because 6000 K blackbody radiation at the wavelength of 10 µm (4348 W m−2 str−1µm−1) is much greater than the 300 K radiation (10 Wm−2 str−1µm−1). Figure 9 shows the time variations in brightness of afterimages of the sun after sunlight input of which the maximum duration is two minutes. Although the brightness becomes larger as the time duration of sunlight input becomes longer, they decrease to around seven counts that are negligible values after ten days. The UMBA does not seem to suffer obvious or permanent damage by the sunlight input of at least two minutes. It is supposed that the afterimages of strong light input are caused by the changes of the physical properties of the bolometer due to the strong heat inputs, and that they will disappear as these physical properties recover with time.

Fig. 9
figure 9

Time variations of luminosity of afterimages after direct sunlight inputs. The time of the sunlight input was varied in steps between 1 second (minimum) to 2 minutes (maximum).

3.4 Temperature control of camera

As stated in the previous section, the UMBA detects not only the thermal radiation from the Venusian atmosphere, but also that from the lens and optical mount. The temperatures of lens and optical mount must be controlled so that their influences on an image are minimized. It has been estimated by calculation, using the nominal transmittance of the lens, that a stability of ±2.3 K is required for the lens temperature, which is almost equivalent to the optical mount temperatures. In order to confirm the temperature range requirement, the flight model of LIR took images of a uniform temperature blackbody under different thermal conditions. Both of the signal outputs for the shutter and the object must be covered simultaneously by the 12 bits dynamic range of the A/D converter. It is concluded that if the optical mount temperature is within 293.5–304.6 K, LIR can take an image of an object whose temperature is as low as 203 K.

4. Summary

LIR developed for Akatsuki, the first Japanese Venus Climate Orbiter, the first lightweight infrared camera to use an uncooled micro-bolometer array as an image sensor. The infrared camera observes the thermal emission from the cloud top of Venus in the wavelength region 8 – 12 µm to obtain two-dimensional distributions of temperature and wind vector at the cloud top, where the temperature is typically as low as 230 K. The flight model of LIR has been manufactured and its performance was confirmed by several tests in a vacuum environment. It has been shown that the target performance of NETD∼0.3 K at an object temperature of ∼230 K is achieved by averaging several tens of images acquired within a few minutes. This NETD corresponds to a height difference at a cloud top of 100 m for an average temperature profile. The camera-case temperature should be stabilized within the temperature range of 293.5–304.6 K in order to take an image of an object having a temperature as low as 203 K. The tolerances of the bolometer array to sunlight, the mechanical environment, and high-energy protons, have been tested and satisfy the requirements. It is also confirmed that the shutter mechanism has enough endurance for the mission life of two years.

LIR has been mounted on Akatsuki which was launched on 21 May, 2010. LIR and the other cameras onboard Akatsuki have taken images of the Earth immediately after the launch. Furthermore, LIR has taken images of deep space for calibrations of the UMBA. However, the Venus orbit insertion maneuver for Akatsuki on December 7, 2010, failed. At present the spacecraft is orbiting the Sun, and it will have a chance to encounter Venus in 5 or 6 years time. JAXA is examining the possibility of conducting an orbit insertion maneuver again at this opportunity.