An essay on $d_0$ in neutron-based strain measurement techniques: equilibrium-based methods for non-destructive $d_0$ estimation

Technical Summary: Equilibrium-Based Methods for Non-Destructive $d_0$ Estimation in Neutron Strain Measurement

The Problem: The Ambiguity of $d_0$
In neutron diffraction-based strain measurement, the calculation of elastic strain ($\epsilon$) relies on the equation $\epsilon = (d - d_0)/d_0$, where $d$ is the measured lattice spacing and $d_0$ is the stress-free reference lattice spacing. While often treated as a material constant, $d_0$ is frequently neither constant nor intrinsic. It varies spatially and directionally due to composition gradients, phase transformations, texture, and residual stresses at the micro- and inter-granular scales (Types II and III stresses).

The fundamental difficulty, termed the "$d_0$ problem," is that diffraction measurements alone cannot distinguish between a change in the reference lattice spacing ($d_0$) and a uniform hydrostatic elastic strain. Mathematically, a set of diffraction measurements is insufficient to uniquely solve for both the strain tensor and $d_0$ simultaneously because the system is rank-deficient; a uniform hydrostatic strain is observationally indistinguishable from a shift in $d_0$. Consequently, determining an appropriate $d_0$ is a critical, non-trivial task. The conventional solution involves destructive testing—cutting the sample into small, stress-relieved coupons to measure $d_0$ directly. However, this is often impractical or impossible for valuable, large, or unique artifacts where destructive sampling is prohibited.

Methodology: A Hierarchy of Equilibrium Constraints
This paper proposes a hierarchy of non-destructive alternatives based on the principle of mechanical equilibrium to estimate $d_0$ without destructive reference samples. The approach leverages the fact that residual stress fields within a sample must satisfy equilibrium equations ($\nabla \cdot \sigma = 0$) and boundary conditions (e.g., traction-free surfaces).

The methodology progresses through three levels of constraint:

1. Force Balance (Global Equilibrium):
   For a planar cross-section within a sample, the net force integrated over the area must be zero (in the absence of external loads). By integrating the measured normal strains across a cross-section, one can derive an integral equation (Equation 15) to solve for a single, constant $d_0$ value that satisfies global force balance. This method assumes isotropic $d_0$ but utilizes the full dataset of strain measurements to resolve the hydrostatic ambiguity.

2. Boundary Conditions (Local Equilibrium):
   On free surfaces where the traction vector is zero ($\sigma \cdot n = 0$), the stress components must vanish. By extrapolating measured strain data to the boundary, one can apply these zero-stress conditions to solve for local $d_0$ values. This effectively adds extra equations to the system of unknowns, allowing for the unique determination of $d_0$ at the boundary, which can then be used to constrain the interior.

3. Point-Wise Equilibrium via Eigenstrain Analysis:
   For more complex scenarios, the paper utilizes the eigenstrain framework. Total strain is decomposed into elastic strain ($\epsilon$) and inelastic eigenstrain ($\epsilon^*$). The authors formulate an inverse problem where a spatially varying $d_0$ distribution and an eigenstrain field are reconstructed simultaneously to fit the observed $d$-spacings.

   • Axisymmetric Systems: For cylindrical geometries, equilibrium reduces to a first-order ordinary differential equation (Equation 18). This allows for the explicit reconstruction of a radial $d_0(r)$ distribution from radial, hoop, and axial lattice spacing measurements, provided the boundary condition at the surface is satisfied.
   • Symmetry Planes and 2D Systems: For planar measurements (e.g., symmetry planes in cubes or cross-sections of blades), the authors construct a basis of eigenstrain fields (using polynomial functions). They solve a regularized least-squares optimization problem to find the coefficients of these basis functions and the parameters of a $d_0(x,y)$ distribution that best fit the data while strictly satisfying equilibrium and boundary conditions. This approach uses the eigenstrain model as a physical regularizer to handle the ill-posed nature of the inverse problem.

Key Results and Case Studies
The paper validates these methods using three distinct examples:

• Additively Manufactured Inconel Cube:
  A Selective Laser Melting (SLM) Inconel cube was measured via neutron diffraction. A stress-relieved "comb" sample cut from an identical cube provided direct $d_0$ measurements for comparison.

  • Results: The $d_0$ values derived from force balance (1.085808 Å) and boundary condition extrapolation (1.085810 Å) were consistent with the direct measurements (average 1.085869 Å) and each other. The point-wise equilibrium reconstruction yielded an average of 1.085792 Å with a small spatial variation. This confirmed that equilibrium-based methods can accurately recover $d_0$ even when direct measurement is possible, serving as a robust validation.

• Ancient Roman Bronze Medical Probes:
  Four ancient bronze probes were analyzed where destructive sampling was impossible.

  • Results: A constant $d_0$ assumption failed to satisfy the surface boundary condition (predicting ~100 MPa radial stress at the surface). A direct application of the axisymmetric differential equation produced a $d_0$ distribution that satisfied equilibrium and boundary conditions but resulted in a net unbalanced force. Finally, the eigenstrain-based reconstruction satisfied both the boundary condition and global force balance simultaneously. The resulting $d_0$ showed a small but measurable variation (range ~3.9 × 10⁻⁴ Å), suggesting that enforcing equilibrium constraints can reveal spatial variations in $d_0$ that constant assumptions miss.

• Ancient Persian Bronze Dagger:
  A bronze dagger (c. 1000 BC) was scanned using a combination of neutron imaging and diffraction.

  • Results: Similar to the probes, a constant $d_0$ was insufficient. The eigenstrain-based reconstruction of a spatially varying $d_0$ (range ~±1.9 × 10⁻⁴ Å) produced a stress field consistent with casting/quenching processes and potential cold-working. The reconstructed $d_0$ showed slightly larger values near the surfaces compared to the core. The residuals between the model and data were within the typical peak-fit uncertainty, demonstrating the feasibility of the approach for non-destructive analysis of unique artifacts.

Significance and Claims
The paper does not claim to prove that $d_0$ definitely varies in these specific samples, acknowledging that the inferred variations are one of many possible solutions consistent with the data. However, the authors assert that their work demonstrates a critical shift in methodology:

1. Non-Destructive Viability: Equilibrium assumptions provide a practical, physical constraint for estimating $d_0$ when destructive reference measurements are entirely inappropriate.
2. Physical Consistency: By treating $d_0$ not as an independent calibration constant but as a variable constrained by the mechanics of the sample, the resulting strain and stress fields are guaranteed to be physically admissible (satisfying equilibrium and boundary conditions).
3. Hierarchy of Constraints: The paper establishes a scalable framework ranging from simple force balances to complex point-wise eigenstrain reconstructions. This hierarchy allows researchers to supplement, test, or replace direct measurements depending on the sample geometry and the availability of symmetry or boundary conditions.

In conclusion, the authors argue that $d_0$ should be reconstructed using the mechanics of the sample whenever possible, moving beyond the simplistic view of it as a fixed material constant. This approach transforms the $d_0$ problem from a calibration hurdle into a solvable inverse problem grounded in the fundamental laws of elasticity.