The Herschel telescope was located outside the cryostat and protected by the sunshade from direct radiation from the Sun. The measured telescope temperature was in the range 83-90 K and showed an annual variation due to the changing heliocentric distance of the Earth and thus L2 (see below). At this temperature, even given a low emissivity, the source contribution is almost always only a small fraction of the telescope background. For comparison, the telescope background flux was of the order of 1000 Jy, while that of Uranus was ∼ 250 Jy and Neptune ∼ 100 Jy. Therefore, a precise characterisation of its behaviour is of critical importance.

Figure 4.1. Temperatures of the primary mirror (M1), cryostat vacumm vessel (CVV) and sun-shield measured from OD40 to OD660. The monotonic increase of temperature from up to OD300 was well correlated to the seasonal temperature variation model.

The telescope background depends primarily on:

Small spatial and temporal temperature gradients are important to the background stability. The requirements on the primary mirror (M1) are:

See "Self-emission" under Section 4.4, “Straylight”

Thermal emission from interstellar dust (known as interstellar cirrus) dominates the FIR Sky Background (FIRSB) at lower Galactic latitudes, while the Cosmic Far-Infrared Background (CFIRB) is more significant towards higher Galactic latitudes, also dominating the confusion noise in the PACS and SPIRE photometric bands. Intrinsically diffuse and unresolved components of the FIRSB are (in descending order of their relative contribution, see [RD4] and [RD5]):

A detailed description of the different components of the FIR background is given in the Herschel Confusion Noise Estimator (HCNE) tool Science Implementation Document ([RD4] and [RD5]). The HCNE tool can be accessed as a standalone service to provide background estimates (see the HSC website for more information), or through the HSpot proposal preparation tool.

While the zodiacal light emission is a major contributor to the sky brightness in the MIR range, it is less important for the FIR and sub-mm wavelengths. Moreover, this emission is quite smooth, lacking fluctuations on arcmin scales (the angular resolution of the ISOPHOT instrument on board ISO). Smaller scale fluctuations, in principle, are likely to exist, but the presence of such structures have not been yet confirmed by the recent observations of the Spitzer Space Telescope.

Confusion noise due to the integrated FIR-sub-mm emission from faint asteroids, individually below the detection limit, has been investigated by Kiss et al. (2006) (see [RD5] and references therein). It has been found that the distribution of asteroids concentrates towards the local anti-solar direction, with a corresponding peak of the confusion noise in the anti-solar point, and an extended cloud is present around the maximum. Seasonal variations are also detected. The confusion noise induced by the cloud of asteroids would only be not negligible in the area around the anti-solar direction, but this area of the sky is closed to Herschel anyway due to the satellite's Sun constraint (see Section 2.3), so the asteroid cloud component is not considered in the HCNE.

Figure 4.2. The brightness of the night sky, excluding the contribution of the extragalactic background (from [RD5], adapted from Leinert et al. 1998, A&A, 127, 1). The spectral ranges covered by the PACS and SPIRE instruments of the Herschel Space Observatory are indicated. Atmospheric contributors, affecting ground-based observation in the optical and NIR, have been also displayed. In the absence of confusion, the far infrared sky is darkest around 350 microns where zodiacal light, CMB and cirrus are all close to their minimum values.

The interstellar medium shows a strong concentration of emission around the Galactic plane; this feature is conspicuous at many wavelengths. However, the cirrus emission is not limited to low Galactic latitudes. It consists of thermal emission of dust in low-density, cool interstellar HI clouds (typically with T≈20K and n≤10²cm^-3), showing a smooth, modified blackbody SED. It is a strong source of emission, and dominates the sky for wavelengths λ>70μm, even at high Galactic latitudes. The cirrus emission is highly structured, and shows a typical filamentary structure.

The main characteristic of the cirrus emission is its spatial structure at a specific wavelength. This is usually described by the spectral index, α,of the power spectrum of the image, averaged over annuli ([RD7]). With this parameter the power spectrum is P=P_o(f / f_o)α, where P is the power at the spatial frequency f and P_o is the power at the spatial frequency f_o. Due to this parameterisation the structure of cirrus is equivalent to that of a fractal.

According to [RD5], cirrus confusion noise can be generally described by the following equation:

σ_cirrus = c₁ × (λ/D)^1-α/2 × B^η

Here σ_cirrus is the confusion noise due to the cirrus component, B is the surface brightness of the field α is the spectral index of the logarithmic power spectrum, averaged in annuli (see [RD5] and references therein), λ is the wavelength of the observation and D is the effective diameter of the telescope's primary mirror. The parameters c₁ and η have to be determined from measurements. This is used within the HCNE to compute the noise due to cirrus emission. Details of the computations are given in [RD5].

In many practical cases, Galactic cirrus confusion noise has been found to be easily parameterised as follows (see for instance [RD6] and references therein):

σ_cirrus ∼ 0.3(λ₁₀₀)²(D_m)^-2.5⟨Bλ⟩^1.5

where σ_cirrus is given in mJy, λ₁₀₀ is the wavelength ratio λ/(100 μm), D_m is the telescope diameter in metres and ⟨Bλ⟩ is the sky brightness in MJy/sr. If we consider fiducial values ⟨B₇₀⟩ = 0.12 MJy/sr and ⟨B₁₆₀⟩ = 1.5 MJy/sr (corresponding to N_HI = 10²⁰ cm^-2) and D_m = 3.5, we get that σ_cirrus(70 μm) = 0.22 μJy and σ_cirrus(160 μm) = 0.08 mJy.