COSINE: Optimizing the production of the X-ray optics for the NewAthena mission #SWI2026

Cosine (https://www.cosine.eu) is responsible for the development of the optics for the NewAthena mission. We do this using the proprietary technology Silicon Pore Optics (SPO).

The NewAthena telescope’s primary mirror assembly will have a diameter of approximately 2.5 meters, made of 600 mirror modules (MM), divided in 15 different configurations, changing with the radius.

Cosine will manufacture the 600 SPO MMs required, and deliver them to an external party for integration into the spacecraft. Each MM has cutting-edge performance requirements and stringent reliability requirements to guarantee their survivability, functionality and performances in all phases of the mission, from ground handling to space operations.

Figure 1: SPO MM manufacturing, from Silicon wafers, through plates and stacks, up to MM integrated in the NewAthena telescope.

The MM manufacturing chain is shown in a simplified form in Figure 1: silicon wafers (commercially available) are cut into plates, the plates are coated, cleaned and stacked, then four stacks are aligned and assembled into a mirror module, which is then integrated in the NewAthena telescope.

The full chain is split in several processes, some entrusted to external suppliers and about 20 controlled and executed directly by cosine. Each process has dedicated equipment, resources, production time, scheduling constraints, and yield. The processes are currently at different technology levels, some are fully automated, others are manual, some require dedicated lengthy configuration and calibration.

Total production time of each MM is of the order of 10 weeks. cosine will have to deliver the 600 MM to the spacecraft integrator at a rate of 1MM/day.

The 600 MM will have to be produced in parallel and the schedule will have to account for production yield, anomalies resolution, maintenance and calibration, resources sharing, equipment occupancy, storage, upper level integration constrains, and other factors to be coordinated with customer and external parties.

To this aim, cosine is developing production and costs models.

The input to these models are the process descriptions and real-life data gathered from ongoing development activities, where cosine is producing several prototype MMs.

The output of the models shall be an estimate of costs and schedule for the production of the 600 MM to be delivered for flight, plus margins, spares, and any item needed for development and qualification activities. The models shall eventually help scheduling all activities, equipment occupancy, and resources allocation.

 

Figure 2 shows an example of scheduling that is expected as result of the planning models. In this example, the process is strongly depending on the MM configuration. The occupancy of multiple machines is schedule to accommodate: development of a new MM configuration, production trials, and actual production for flight.

 

 

 

 

Figure 2: Example of process scheduling.

 

The models can be initially deterministic and will have to be extended to a statistical approach to account for all uncertainties, risks, anomalies and production yield.

In both cases the models shall be able to accommodate changes such as:

• addition or removal of whole processes from manufacturing chain,
• process parameters (e.g. production time and yield),
• high level requirements (e.g. required production rates, performances, or yield),
• resources (e.g. reduced availability of equipment or man-power),
• costs (e.g. variations of materials costs).

The model shall take into account that the technology and manufacturing processes are still being developed and such changes will be introduced up to and including the flight production, when some development activities will still run in parallel.

The challenge of these models is to address the dependencies between processes which have different complexity, throughput, production time, and yield, and understand how yield and anomalies of a given process can affect the whole chain, as well as identify suitable key performance factors allowing monitoring and continuously improving of production quality and reliability, while keeping costs and delivery schedule under control.

We are looking for a mathematical and statistical framework in which the complexity of the process can be described, so that realistic cost and time estimates can be computed.