The IceCube Neutrino Observatory is a cubic-kilometer neutrino detector at the geographic South Pole [1]. Reconstruction of the direction, energy and flavor of incident neutrinos relies on the detection of Cherenkov radiation emitted by charged particles produced by neutrino interactions with a nucleon in the ice or bedrock. The photons are collected by Digital Optical Modules (DOMs), installed in boreholes drilled into the ice at depths ranging between 1450 m and 2450 m. The IceCube Upgrade will consist of nearly 800 new optical modules on 7 strings. There are two main optical module designs for the Upgrade: the mDOM [2] (figure 1), featuring 24 Photomultiplier Tubes (PMTs) of 3 diameter, yielding an almost homogeneous angular coverage, and the DEgg [3] (figure 1), with two 8 PMTs opposite each other. The Long Optical Module (LOM), developed from the multi-PMT Upgrade modules, has a reduced diameter (12 ) to enable the module to fit in a narrower hole, saving time and fuel costs for drilling. Multiple 4 PMTs have been used to maximize effective area with fewer channels. The waveform processing has been
shifted to the PMT base, integrating digitization and high voltage (HV) generation, and reducing power consumption. Two variations of this design are currently under development (figure 1), with 18 and 16 PMTs. In the near term, the goal is to build ten modules of each type, and deploy 12 modules for the Upgrade. These will be developed into a single baseline design for Gen2, which will have 9600 modules distributed across 120 strings [4].
Each PMT is coupled to the glass vessel with optical gel, which is a transparent addition-curing silicone. The gel, while liquid, could be injected into a clear plastic 'shell' around each PMT photocathode (in situ method (figure 2)). Another alternative is investigated for the LOM, where gel is moulded onto the PMT photocathode as 'pads' (figure 3), which are then coupled to the vessel. This approach is more economical as the pad moulds can be reused. Our implementation of the gel pad is to cast a conical pad directly onto the PMT photocathode. A cavity is formed when the rim of the base makes contact with the curved pressure vessel (figure 4). Silicone caulk is used to seal the rim of the pad to the pressure vessel. The cavity is then filled with liquid gel, and degassed to remove any bubbles. The PMT-and-pad assembly must be precisely oriented for proper interfacing, and the PMT must be pushed against the vessel by the support structure. Some options to exert the outward pressure are springs or an inflatable collar around the PMT neck.
The use of conical gel pads provides photon capture efficiency comparable to metallic reflectors, due to total internal reflection at the surface of the pad (figure 5). This was experimentally verified using a gel pad mounted on a 3 PMT, and counting the number of photons captured when a spot beam is moved across the photocathode area [5]. GEANT4 [6] studies were also performed to optimise the pad shape for greater photon capture efficiency. The probability of the photons arriving at the PMT photocathode was evaluated as a function of the gel pad cone opening angle, and found to be largest at opening angles of 60 • -80 • .
The 4 PMTs used are assumed to have the optical response of the 3 PMTs used in the mDOM, scaled to account for the greater photocathode area.
GEANT4 simulations also enabled the comparison of the module effective area and angular acceptance with other optical modules (figure 6). These yielded an increase in Cherenkov spectrumweighted effective area of 3.2-4.2 times for the 16-and 18-PMT LOMs respectively, relative to the IceCube Gen1 DOM, at photon wavelength 400 nm [5].
The electronics design for the LOM builds upon the mDOM design, adding data acquisition (DAQ) functionality to the mDOM MicroBase, which has the PMT biasing circuit. The new base, called the Waveform MicroBase, is shaped to optimise board surface area to fit the DAQ components in the space between PMTs.
Following the existing IceCube Upgrade designs, a central processor will provide common data management and control. Communications to the hub surface computers via the in-ice cables (figure 7) is managed by the Ice Communications Module (ICM) with a dedicated FPGA. Ribbon cables carry UART communication, time synchronization clocks, and power to the PMT bases. To facilitate a multi-level detector trigger, hit data are temporarily retained in a flash memory chip [7]. The Waveform MicroBase (figure 8) has been optimized for low power consumption and wide dynamic range. We use a 5 mW Cockcroft-Walton generator for HV bias. To accommodate IceCube neutrino events with PMT signals of varying intensity, the base includes two analog channels, one for the anode and another connected to a dynode (Dy8) with lower gain. These signals are digitized continuously in a 2-channel ADC at 60MSPS, and captured in a low power FPGA. For an inter-string spacing of 240 m, as anticipated for IceCube Gen2, photon arrival time distributions have widths >25 nsec, and the choice of 60MSPS is a suitable compromise between power consumption and time resolution. The anode channel remains linear for intensities up to 50PE/25 nsec. A delay line module in the FPGA records the leading edge time with resolution about 1 nsec.