A part comes back from the field with a discoloured contact, or a batch of incoming components does not look like the last one, and somebody has to book an instrument. Both candidates are X-ray methods, both are non-destructive, and both produce an image, so both get proposed for the same problem. They answer different questions. One tells you where the material is and where it is missing. The other tells you what it is made of.
Put that way the choice is usually obvious. The difficulty is that nobody arrives with it put that way. They arrive with a symptom, and symptoms do not announce whether their cause is geometric or chemical.
Micro-XRF and CT are not measuring the same thing
CT builds a three-dimensional reconstruction from X-ray projections taken while the sample rotates. What it maps is attenuation, which follows density and thickness. That is why micro-CT is strong on porosity, voids, cracks, internal geometry and assembly faults. Those are all places where material is present or absent, and absence stands out.
Micro-XRF works from a different effect. The sample is illuminated with X-rays, atoms in it fluoresce, and each element emits at its own characteristic energy. The detector counts those energies, so the reading is an elemental fingerprint of whatever the beam sits on. Focus that beam to a small spot, step it across the sample, and the output is a map of which elements are present and where. Nothing is dissolved or sectioned, so the object goes back into service.
The questions that only micro-XRF can answer
Four questions come up repeatedly, and no scan settles them. Is this the alloy we ordered? Incoming goods can be verified with no sample preparation. How thick is the plating, and is it the right composition? Layer thickness and composition are both measured without sectioning the part, which matters when the check meant to release it would destroy it. What is this contaminant? An inclusion can be named by element and located at once. Is the distribution uniform? Mapping shows how a component is spread across a surface, not one averaged number.
The instrument behind most of this work is Bruker's micro X-ray fluorescence line, a range of spectrometers rather than one fixed configuration. Merkel Technologies represents Bruker Nano Analytics in Israel and handles installation, training and service locally.
When the sample itself decides the method
Bulk analytical methods want a prepared, homogeneous, flat specimen. Much real work does not look like that. An irregular casting. An inclusion inside a weld. One component on a populated board. An object nobody will let you cut. Micro-XRF is the method of choice for these cases: non-homogeneous or irregularly shaped samples, small objects and inclusions, measured largely as they are.
That is why it turns up in fields with little else in common, from electronics failure analysis to geology and heritage conservation. The object matters more than the measurement, and it has to survive the answer.
Where micro-XRF is the wrong tool
Depth is the first limit and the one most often missed. The fluorescence that reaches the detector comes from at or near the surface. A void inside a casting, a crack under a solder joint or a separation inside a laminate will not appear on any map, because the signal never leaves the interior. That is CT territory, and at high resolution it belongs to a laboratory system such as the RX Solutions EasyTom line.
The same surface sensitivity sets a subtler trap. A reading on a coated, oxidised or simply dirty surface describes the surface, not always the part. In materials verification that distinction separates a useful result from a misleading one, so decide in advance which layer the answer describes.
Two further limits deserve stating plainly. XRF covers a wide elemental range but is not a universal method, and light elements at trace levels belong elsewhere. It also reports elements, not chemistry: it can show iron and oxygen in a spot without naming the oxide they formed. Compounds and crystal structure need another instrument.
Some problems are mechanical, not chemical. A disbonded adhesive joint or a delamination in a composite is not a composition question at all, and is answered faster by non-contact ultrasonic inspection than by either XRF or CT.
CT has its own boundary, the mirror image of this one. A scan shows an inclusion as a bright speck of higher attenuation, but attenuation does not name an element. Materials of similar density are hard to separate, and no scan alone says whether a speck is steel, solder or a carbide.
Running CT and micro-XRF in the right order
In failure work the productive pattern is not choosing between the two but sequencing them. CT goes first and locates the anomaly: this inclusion, at this depth, this close to the weld. Micro-XRF goes second, once the region is reachable, and names it: this element, therefore that process step. One answers where, the other what, and a component failure investigation needs both before anyone can say why.
Reversing that order wastes time, because mapping a whole surface in the hope of meeting the defect is slow when a scan would have pointed straight at it. The two belong to the same material analysis and non-destructive testing toolkit rather than competing for one budget line.
Buy the answer, not the instrument
Most laboratories and QA groups do not need to own both. They need to know, before the purchase conversation starts, which question dominates their week. A group verifying materials, checking coatings and identifying contamination will get more from a micro-XRF spectrometer than from a scanner. A group chasing porosity and assembly faults will reach the opposite conclusion.
The honest test is to put real samples in front of both and see which produces something you can act on. Merkel Technologies runs demonstrations and proof-of-concept analysis on customers' own samples in Israel, and the useful result is often finding that only one was needed.