APT · Correlative microscopy

Correlative APT and TEM on the same needle

Image one needle with electrons first (TEM or TKD, which show crystal structure), then take it apart atom by atom in an atom probe (APT, which names each atom): structure and chemistry from the same spot.

What happens to the needle in a correlative TEM and APT study?

One needle, five steps, in the order used by Herbig et al. (2015). FIB: a focused gallium-ion beam that cuts the sample.

Schematic, not to scale; the zoom box in step 2 is to scale.
This step gives
Ga damage (Si)
Carbon cap

Try it: switch 30 kV to 2 kV: the coral damage layer shrinks from about 40 nm to under 1 nm.

How does a TEM image of the tip change the atom probe reconstruction?

The atom map's depth is calculated from the tip radius and cone half angle (shank angle). Left: needle with flat test layers; right: where the map puts them.

Model: a sphere on a cone. Test layers are the page's choice; widths are schematic.

Needle, measured in TEM

Used in the reconstruction

Last layer, true
In the atom map
Depth error
Width there

Try it: press Copy: the coral layers sit on the grey ones and the depth error reads 0%.

What does TEM show that APT misses, and the other way round?

The same needle top, seen by each technique (solute: the minor element).

Schematic. Based on Herbig et al. (2014, 2015).
TEM
APT

Try it: pick Lattice planes and switch to APT: the fringes vanish.

What to take away

One needle, two views

TEM gives the structure. APT names each atom by its flight time to the detector.

TEM first, APT last

APT destroys the part it measures, so that part is imaged first.

The tip outline feeds the map

Radius and shank angle set the depth scale. In this model, ignoring the shank angle stretches depth.

Measured is not always best

Herbig et al. (2015) measured 50 nm and 10°; the best map used 30 nm and 7°.

More detail: the equations, the numbers and the limits

The shank-angle reconstruction

In the standard protocol each detected ion moves the virtual surface down by dz = Ω/(η SA), with SA = 2πR2(1 − cos θD): Ω is the atomic volume, η the detection efficiency, θD the angular half field of view (Gault et al. 2011). In the shank-angle protocol, where the tip is taken as a sphere on a truncated cone (Larson, Gault et al. 2013), the radius grows with depth as R = R0 + w z, with w = sin α/(1 − sin α) and α the half shank angle (De Geuser and Gault 2017). Sideways sizes scale with R, because the magnification is about L/(ξR) (Hatzoglou et al. 2025).

What the page computes

The number of atoms collected down to depth z is proportional to the integral of R2, so N ∝ [(R0 + wz)3 − R03]/3w (or R02z when α = 0). The page finds N for each true layer with the TEM values, then the depth that gives the same N with the values used. Ω, η and θD are taken equal in both, so they cancel. The width ratio is R used over R true at each layer. In this simple model, 30 nm and 7° would not match a 50 nm, 10° tip; Herbig et al. explain their result by effects the model leaves out.

The other way to get R

In voltage mode the radius comes from the applied voltage: R = V/(k F), with F the evaporation field and k the field factor; typical k values range from 2 to 8 (Gault et al. 2021). Larson, Gault et al. (2013) give the radius either from the voltage or from the specimen geometry.

Numbers on the page

Ga damage: in {100} Si, Ga+ at 30 keV goes about 40 nm deep; clean-up at 5 or 2 keV leaves about 5 nm and under 1 nm of damaged Si (Thompson et al. 2007). Other materials differ.

TEM: Herbig et al. (2015) imaged needles at 200 kV; their needle was under 60 nm thick in the first 400 nm along the axis. They explain the 30 nm against 50 nm by the first atoms not being used in the reconstruction, and the 7° against 10° by the limited field of view of APT.

TKD: Babinsky et al. (2014) placed a grain boundary in the first 220 nm of a molybdenum needle. Best results came at 30 kV with the surface tilted −45° to −35°, and orientations could be read up to about 200 nm thickness of molybdenum.

Limits

The model is a sphere on a cone with a fixed field of view. Real tips are often not hemispherical, and different phases can have very different evaporation fields (Larson, Gault et al. 2013); this model leaves that out. Tip radius or shank angle from correlative microscopy do not necessarily give the best reconstruction (Larson, Gault et al. 2013; De Geuser and Gault 2017).

On this site: APT reconstruction and resolution · APT instrument and prep · FIB prep artifacts · EBSD lift-out planner · APT and short-range order · 4D-STEM basics · Grain boundaries (CSL)

Questions people ask

Can an atom probe needle go into a TEM?

Yes. The needle can sit on a TEM grid of the right size (Gault et al. 2021). Herbig et al. (2015) adapted a single-tilt TEM holder so the same mount fits the FIB, the TEM and the atom probe. The top must be thin enough for the beam; theirs was under 60 nm thick in the first 400 nm.

Does the electron beam contaminate the tip before APT?

Molecules picked up in air can build carbon caps on the tip under an electron beam. Herbig et al. (2015) report that oxygen plasma cleaning avoids or removes this. They never saw a change in the carbon content measured by APT after imaging in a 200 kV TEM.

Why use TKD on an atom probe tip?

Transmission Kikuchi diffraction (TKD) gives the crystal orientation of thin regions. Babinsky et al. (2014) used it with FIB to place a grain boundary in the first 220 nm of the tip, faster than with TEM. It also gives the crystallography that helps read the APT data.

Why doesn't the radius from TEM give the best reconstruction?

The simple reconstruction assumes a hemisphere on a cone, which a real tip is not. In Herbig et al. (2015) the TEM outline gave 50 nm and 10°, but 30 nm and 7° matched the image best, because the first atoms were left out and the field of view is limited. Larson, Gault et al. (2013) also note that radius or shank angle from correlative microscopy do not necessarily give the best reconstruction.

Why finish the FIB needle at low kV?

Gallium at 30 keV damages a surface layer. In {100} Si, Thompson et al. (2007) found about 40 nm of Ga penetration at 30 keV, about 5 nm of damage after a 5 keV clean-up and under 1 nm after 2 keV.

What does combining TEM and APT on one tip give?

Herbig et al. (2014) measured the five crystallographic parameters of grain boundaries by TEM and their carbon content by APT on the same tips. In ferrite, carbon segregation at low-angle boundaries rose roughly linearly with misorientation angle. For high-angle boundaries there was no general trend with misorientation angle.

References

Show the 10 references
  1. D. J. Larson, T. J. Prosa, R. M. Ulfig, B. P. Geiser and T. F. Kelly, Local Electrode Atom Probe Tomography: A User's Guide, Springer (2013): chapter 2, Specimen Preparation (pp. 25 to 53), and chapter 5, Data Processing and Reconstruction (pp. 109 to 162). doi:10.1007/978-1-4614-8721-0
  2. K. Thompson, D. Lawrence, D. J. Larson, J. D. Olson, T. F. Kelly and B. Gorman, In situ site-specific specimen preparation for atom probe tomography, Ultramicroscopy 107, 131 to 139 (2007). doi:10.1016/j.ultramic.2006.06.008
  3. M. Herbig, P. Choi and D. Raabe, Combining structural and chemical information at the nanometer scale by correlative transmission electron microscopy and atom probe tomography, Ultramicroscopy 153, 32 to 39 (2015). doi:10.1016/j.ultramic.2015.02.003
  4. M. Herbig, D. Raabe, Y. J. Li, P. Choi, S. Zaefferer and S. Goto, Atomic-scale quantification of grain boundary segregation in nanocrystalline material, Physical Review Letters 112, 126103 (2014). doi:10.1103/PhysRevLett.112.126103
  5. K. Babinsky, R. De Kloe, H. Clemens and S. Primig, A novel approach for site-specific atom probe specimen preparation by focused ion beam and transmission electron backscatter diffraction, Ultramicroscopy 144, 9 to 18 (2014). doi:10.1016/j.ultramic.2014.04.003
  6. B. Gault, S. T. Loi, V. J. Araullo-Peters, L. T. Stephenson, M. P. Moody, S. L. Shrestha, R. K. W. Marceau, L. Yao, J. M. Cairney and S. P. Ringer, Dynamic reconstruction for atom probe tomography, Ultramicroscopy 111, 1619 to 1624 (2011). arXiv:1510.02849
  7. F. De Geuser and B. Gault, Reflections on the projection of ions in atom probe tomography, Microscopy and Microanalysis 23, 238 to 246 (2017). doi:10.1017/S1431927616012721 (arXiv:1606.08064)
  8. D. J. Larson, B. Gault, B. P. Geiser, F. De Geuser and F. Vurpillot, Atom probe tomography spatial reconstruction: status and directions, Current Opinion in Solid State and Materials Science 17, 236 to 247 (2013). doi:10.1016/j.cossms.2013.09.002
  9. C. Hatzoglou, P. Kontis, H. Khanchandani, G. Da Costa, B. Gault and F. Vurpillot, Effect of dynamic reconstruction on particle size and morphology in atom probe tomography, Microscopy and Microanalysis 31(2), ozaf006 (2025).
  10. B. Gault, A. Chiaramonti, O. Cojocaru-Mirédin, P. Stender, R. Dubosq, C. Freysoldt, S. K. Makineni, T. Li, M. Moody and J. M. Cairney, Atom probe tomography, Nature Reviews Methods Primers (2021). doi:10.1038/s43586-021-00047-w
Cite this page: Tripathy, Manisha. “Correlative APT and TEM Lab.” untethered atom, 2026, https://untetheredatom.com/apt/correlative-apt-tem.
BibTeX 
@misc{tripathy2026correlativeapttem,
  author = {Tripathy, Manisha},
  title  = {Correlative APT and TEM Lab},
  year   = {2026},
  howpublished = {\url{https://untetheredatom.com/apt/correlative-apt-tem}},
  note   = {Interactive web tool}
}
Last updated 28 September 2026.