Cryo-electron microscopy: why cooling must be constant, not powerful

01Little cooling, high requirement

Cryo-electron microscopes make individual proteins visible. For this, the sample is flash-frozen and kept below −170 °C with liquid nitrogen; from tens of thousands of individual images, a three-dimensional model is created with a resolution of a few ångströms, i.e. fractions of a nanometre. A measurement series often runs unattended for hours to days – an interruption can render the entire series worthless.

02The fallacy: cooling capacity instead of constancy

For building services engineering, the task initially looks small: per instrument, only a few kilowatts have to be removed to the cooling water, less than for some server cabinets. The magnetic lenses of the microscope, however, are current-carrying coils whose optical properties depend on their temperature. The cooling water therefore does not just keep the instrument cool, it keeps the optics stable.

03The principle: thermal expansion versus resolution

Metal expands with temperature. A simplified estimate shows the order of magnitude: half a metre of steel becomes around 0.6 micrometres longer when heated by 0.1 K – about three thousand times the target resolution.

Manufacturers therefore specify tight limits for the stability of cooling water and room temperature, often as a permissible change within a time window. For the room air, published facility descriptions state, for example, less than 0.8 K fluctuation over 24 hours.

04Inertia is cheaper than control quality

Whether a tolerance is formulated as an instantaneous band or as drift over a time window makes a big difference. Slow drift can be managed with inertia: a well-mixed buffer volume in the instrument circuit acts like a low-pass filter that smooths out rapid fluctuations from the chilled water network. A plate heat exchanger alone hardly achieves this; it separates systems, but does not dampen.

−92 %less temperature fluctuation at the instrument in the fictitious example with a time constant of 20 minutes

In addition, there is a separate control stage for each instrument, so that faults or the shutdown of one microscope do not affect the others. The network temperature is often overlooked: if the flow temperature of the chilled water network is only slightly below the instrument’s set point, more water has to be mixed in, the control valve works within a usable stroke range and control becomes smoother. Colder is not better here.

05What you should clarify for your project

• Obtain the installation requirements of each instrument early: heat dissipation to water and room, flow temperature, stability, water quality
• For every tolerance, check whether a band or drift over time is meant
• Provide temperature sensors at the instrument inlet with suitable measuring accuracy and recording – as evidence at acceptance
• Design the pipe network so that later operation via dedicated chiller units remains possible
• Plan for redundancy, maintenance windows, vibration and magnetic stray fields from the outset

In this way, a supposedly small cooling task becomes a system that enables undisturbed measurements over days.

Takeaway

When cooling cryo-electron microscopes, constancy is the task, not capacity – and constancy is achieved most economically with inertia rather than with ever faster control loops.

SEDC AG · Article from our internal training series “Technik Praxis”, abridged for clients and users.