NIEL Conversion Tool

Preliminary analysis aid. Check the result against your own analysis and, where it matters, a test on your part. Full disclaimer.

1. What the tool calculates

Proton displacement damage bookkeeping. Non ionizing energy loss (NIEL) at any proton energy for ten elements and for compounds by Bragg's rule; the fluence at one energy that does the same displacement damage as a fluence at another; displacement damage dose (DDD) from fluence and energy; and the 1 MeV neutron equivalent fluence. Six tabs: NIEL Lookup, Fluence Converter, DDD Calculator, 1 MeV n_eq, NIEL Plot (log log, several materials at once), Data Table (the raw Jun et al. values).

NIEL is the displacement damage counterpart of LET: energy per unit path length that goes into displacing atoms rather than ionizing them. It is a property of the material and the particle, not of the device. Device response enters through a damage coefficient measured on the part. What makes NIEL useful is the NIEL hypothesis: displacement damage scales with total non ionizing energy deposited regardless of particle type or energy. It is an engineering approximation with known failures (section 7), and it is the basis of everything on this page.

2. Inputs

• Material. Si, C, Al, P, Ga, Ge, As, In, Cu or Se, or a compound: presets GaAs, InP, SiC, SiGe, InGaAs, CIGS (CuInGaSe₂), or any formula you type as elements with stoichiometric coefficients.
• Proton energy, MeV, within the tabulated range for the material (about 0.2 keV to 1000 MeV for silicon; the low end varies by element).
• Fluence, p/cm2, and for the converter a source energy and a target energy.
• NIEL units: MeV cm2/g, keV cm2/g or MeV mb.

3. Calculation

Data

All NIEL values are from Jun et al. [1]: proton NIEL for C, Al, Si, P, Cu, Ga, Ge, As, Se and In from below 1 keV to 1000 MeV. Coulomb scattering on a ZBL screened potential (relativistic above 50 MeV) dominates below about 10 MeV; nuclear elastic and nonelastic reactions, computed with MCNPX in a thin-target approximation, are generally the larger share above 30 to 50 MeV; the Lindhard partition separates the displacing and ionizing shares of the recoil energy. Every element carries every row of the paper's Table I (silicon 36 rows); the table was re-transcribed against the published paper on 29 September 2026, which corrected the silicon rows from 3 keV to 0.7 MeV and larger errors in the carbon, phosphorus and germanium columns of the earlier copy.

Interpolation

Log log between adjacent tabulated points, equivalent to a local power law. No extrapolation outside the table; the tool says so rather than guessing.

Compounds: Bragg's rule

Mass weighted sum of the elemental values at the same energy, with w_i from the formula's atom counts n_i and atomic masses M_i.

Fluence conversion

Used to turn a mission spectrum into a monoenergetic test fluence, to put two laboratories' tests at different energies on one axis, or to compare with a published result at another energy.

Displacement damage dose

Parameters governed by minority carrier lifetime (bipolar gain, solar cell short circuit current, detector dark current) degrade in proportion to DDD through a damage coefficient K measured on the device. Measure K at any convenient proton energy, compute the mission DDD, and the same K gives the on orbit degradation; that is the point of the whole framework.

1 MeV neutron equivalent

The reference is the silicon displacement kerma of ASTM E722 [5]. Converted with N_A/A the 95 MeV mb figure gives 2.04e-3 MeV cm2/g; the community uses values from 1.95e-3 to 2.04e-3 depending on the kerma calculation, and this tool uses 1.95e-3, so its 1 MeV equivalent fluence is about 4.5 percent higher than the ASTM reference value gives (multiply by 0.957 to convert). The difference is under 5 percent, generally well inside the uncertainty of a displacement damage prediction. The ratio NIEL_p(E) / NIEL(1 MeV n, Si) is the hardness factor. The convention is always referenced to silicon: with silicon selected you get the 1 MeV n_eq on this tool's reference, about 4.5 percent above the value the ASTM reference gives; with another material you get a damage normalization that compares proton damage in that material to the silicon reference, which is useful but is not the ASTM E722 quantity.

4. Units

MeV mb is the per atom damage energy cross section, the native output of the Monte Carlo and the form in Jun et al. Table I. MeV cm2/g is the mass form used for DDD. For compounds use the molecular mass in place of A.

5. Assumptions

• The NIEL hypothesis: damage is proportional to non ionizing energy deposited, whatever produced it.
• Bragg additivity: the compound is a mechanical mixture of its elements. Nuclear interactions at proton energies are far above chemical binding energies, so this is a good approximation where nuclear reactions dominate; Jun et al. checked it against a direct GaAs calculation with good agreement. It ignores differences in threshold displacement energy, crystal structure and defect formation between the compound and its elements, and it degrades at low energies where Coulomb scattering dominates.
• Initial damage only: no annealing, no injection or temperature dependence.

6. Implementation details

• Ten elements at the Table I energies (silicon 36 rows); compounds computed at request time from the elemental values by Bragg's rule.
• Out of range energies return a message, not a number.
• The plot tab overlays any number of materials on log log axes; the data table shows the raw values.
• DDD is shown in MeV/g with scientific notation for large values.

7. Limitations

• Energy range. Below about 1 keV, Coulomb scattering dominates, NIEL rises steeply and depends on the displacement threshold energy; use the low end with care. Above 1000 MeV use NEMO [2] or Geant4 directly.
• Where the NIEL hypothesis fails [4], [6]. Clustered defects from high energy recoils and isolated point defects from Coulomb scattering have different electrical properties, so two exposures with equal DDD but different cluster to point ratios can degrade a device differently; this matters most when comparing protons with electrons or widely different proton energies. Compound semiconductors (GaAs, InP) show sublinear or superlinear behavior for some parameters. Annealing at room temperature reduces damage over time for some silicon defects. Some devices show a dose rate dependence analogous to ELDRS. None of these is in the arithmetic here.
• Protons only. No electron NIEL (solar cells in the belts need it), no neutron or heavy ion NIEL; for heavy ions see Messenger et al. [3].
• The damage coefficient is the real uncertainty. K varies between manufacturers, lots and wafers. A DDD tells you how much displacing energy went in; what the part does with it comes from a test on that part. For anything that matters, measure K under conditions representative of the mission and use this tool to move between energies.

8. Caveats on fluence times NIEL as a damage metric

The displacement damage dose this tool computes, fluence times NIEL, is a good predictor of device degradation for protons across energies and for neutrons, within the scatter of the measurements that established it: protons, deuterons, helium ions and fission neutrons on one line for bipolar gain [8], and protons across energies for silicon, gallium arsenide and indium phosphide devices [4]. Six conditions generally have to hold for the number to mean what it appears to mean.

Electrons are not protons, and the difference depends on the device. In n-type silicon and in bipolar transistors of either base type the electron damage coefficient is linear in NIEL and sits on the proton line [4], [13]; in p-type silicon solar cells it is quadratic, about ten times below the n-type value at 1 MeV [4], so an electron fluence converted with electron NIEL alone overstates solar-cell damage relative to a proton fluence of the same DDD. Solar cell work therefore raises the electron NIEL to a power fitted to the device, between 1 and 3 across metrics (2.4 for silicon maximum power with a 21 eV threshold, often quoted near 1.7) before adding electron and proton contributions [11]. This tool carries no electron table; the exponent, if one applies, must be applied by hand.

The dose behind shielding is not the dose outside it. The proton spectrum hardens with depth, so the DDD per unit shield thickness falls slowly [9]; the slowed-down protons that arrive below 0.1 MeV can still carry a third of the dose behind a thin coverglass [9]; and nuclear interactions along the protons' path in the shield produce secondary neutrons whose displacement damage overtakes that of the surviving primaries beyond about 10 to 16 g/cm2 of aluminum for the most energetic protons, and much sooner in high-Z shields (the secondary pions, deuterons and alphas add at most 5 percent) [10]. A DDD computed from the external spectrum with this tool is a starting point for a transport calculation, not its result.

The particle has to outrange the active volume. When the incident protons slow down or stop inside the active region the damage per proton peaks where the track ends and then falls: silicon solar cells depart from NIEL below about 8 MeV for an 80 um active depth, gallium arsenide cells below 0.3 to 0.4 MeV for half a micrometer [12], [14]. An adjusted NIEL integrated along the track in the active region restores the correlation [14]; this tool's table is for protons that pass through.

NIEL scaling assumes the cascade stays inside the active volume. When the range of the energetic recoils exceeds the active layer thickness, as in thin films and some detector structures, part of the damage energy leaves and NIEL overestimates the damage [12]. Heavy ions have a published NIEL (iron at 1 GeV, some 2 to 3 MeV cm2/g in silicon [3]) and their damage follows it where the recoils are contained; this tool carries no heavy-ion table.

Point defects and clusters are not interchangeable. Per unit damage energy, generation centers in the cluster regime are tens of times more effective than isolated defects [6], which is the electron case above; two exposures at equal DDD with very different recoil spectra can degrade a device differently, and the effective NIEL of [11], which weights the recoil spectrum by the defects that survive the cascade, brings the electron and proton dark-current data onto one line. Among protons and neutrons of every energy, whose recoils all populate the cluster regime, the mean damage per unit NIEL is the same [8], [6].

Finally, the partition function and the displacement threshold that every NIEL table rests on carry their own uncertainty, about 15 percent for the silicon damage metric from the partition function alone [15], and a threshold choice between 13 and 30 eV moves proton NIEL by about 15 percent and electron NIEL below 1 MeV by 40 percent or more [2]; the values here are a calculation with stated assumptions, not a measurement. The interactive model on the tool page shows each of these points with a lattice, a recoil spectrum and six devices.

9. References

Back to the NIEL Tool