Diffraction, microscopy and complementary structure evidence

Different structural measurements observe different features

Diffraction records wave interference related to a sample’s spatial organization. Microscopy forms images through instrument- and specimen-dependent contrast. Elemental analysis constrains composition; molecular spectra can constrain environments and connectivity. They provide complementary evidence, rather than one universally interchangeable kind of “structure picture.”

A diffraction feature need not determine every atomic coordinate, and a calibrated image length need not establish chemical identity. Specimen preparation, spatial or ensemble averaging, disorder, instrument response and the inference model matter. A conventional periodic-crystal calculation should not be presented as a definition encompassing every ordered material.

For example, X-ray absorption spectroscopy can provide element-associated local structural or electronic constraints under its measurement model. It adds a different question to diffraction or imaging; an instrument name does not settle a complete oxidation-state, composition or three-dimensional structure assignment.

A reported 2θ must be halved in the selected Bragg relation

For a conventional periodic-plane model, Bragg’s relation is nλ=2d sinθ, where n is the assigned diffraction order, λ wavelength, d plane spacing and θ the corresponding Bragg angle. An instrument or report can specify the scattering angle 2θ instead. The equation requires θ, not the unhalved reported value.

Give invented wavelength 0.154 nm, reported 2θ=40.0° and first order n=1. Then θ=20.0° and d=0.154/(2 sin20°)=0.22513294 nm. Using 40.0° in the sine would instead give 0.11979073 nm, a different result caused by the convention error.

One selected spacing does not supply full indexing, intensities, atomic positions or a refined crystal structure. Wavelength, angle calibration, peak assignment, order and suitable structural model require evidence in a real measurement. The numbers here are assigned teaching inputs, not an observed diffraction peak.

Intensity can leave different phase descriptions unresolved

Use an original one-dimensional coherent model with two identical real unit-amplitude point scatterers at positions 0 and d. At reciprocal coordinate q, the complex amplitude is F(q)=1+exp(iqd), and normalized intensity is I=|F|²=2+2cos(qd). The product qd is a dimensionless phase.

At phases 0, π/2, π, 3π/2 and 2π, intensities are 4, 2, 0, 2 and 4. At phase π the two amplitudes cancel. This follows from both their relative phase and their stipulated identical amplitudes, rather than from either point being absent.

Translating both points by a common amount multiplies F by a unit-magnitude phase factor and leaves I unchanged. Reflecting the positions in this identical-real-scatterer model also preserves intensity. Thus an intensity observation does not retain every phase distinction. The example is not a full diffraction refinement, anomalous-scattering model or powder pattern.

Changing separation d changes the reciprocal coordinate q of the first positive zero: qd=π. Doubling d halves that q. The plot uses phase qd as its coordinate, so its mathematical shape stays the same while the conversion back to q depends on d.

Pixel scale is different from spatial resolution

A scale conversion relates image pixels to a stated length under appropriate calibration. Give hypothetical calibration length 20.0 nm spanning 50 pixels, so pixel pitch s=0.400 nm per pixel. A selected feature spanning N_p=125 pixels has inferred length L=sN_p=50.0 nm. Pixel pitch is not itself spatial resolution or proof of accuracy.

Assign standard uncertainty u_s=0.00800 nm per pixel and feature-count uncertainty u_N=2.00 pixels, with zero covariance. In a first-order independent-input model, u_L²=N_p²u_s²+s²u_N². The contributions are 1.00 and 0.800 nm, giving combined standard uncertainty √(1.00²+0.800²)=1.28062485 nm.

These are given uncertainty components, not a complete image-measurement uncertainty evaluation. Segmentation, distortion, specimen preparation, calibration correlation and other effects are omitted. If the entire image is represented at twice the pixel count with pitch and uncertainties transformed consistently, inferred physical length does not double.

Combine constraints within a named candidate universe

Suppose the only candidates admitted to an original exercise are propanal and propanone. Both have neutral formula C₃H₆O and a carbonyl group. The simple neutral closed-shell C/H/O unsaturation count is DBE=(2C+2−H)/2=1, where DBE denotes the degree of unsaturation. The formula and carbonyl constraint therefore retain both candidates.

Now stipulate independently verified evidence that all six hydrogens belong to one equivalent set in a single species under the assumed ordinary achiral solution conditions with fast methyl rotation. Within this two-candidate universe, propanone satisfies that constraint while propanal does not. Formula, a carbonyl indicator and verified equivalence answer different parts of the comparison.

The inference assumes the supplied assignment is valid and excludes overlap, exchange, invisible signals and additional candidates. It is not universal identification from a single observed line, nor an actual measured spectrum. Removing the equivalence constraint restores both candidates; repeated use of formula alone does not replace missing independent information.

State what each observation establishes and what remains open

A structural result should distinguish observed quantities from model-derived features. A plane spacing, an intensity pattern, a pixel-calibrated feature and a molecular candidate each carry different assumptions and uncertainty. Attractive visual detail cannot supply missing chemical discrimination.

Use complementary observations to address explicit alternatives, and preserve raw evidence, calibration and inference limits. The teaching figures and examples here are original mathematical or supplied-data models; no source diffraction image, microscopy pixels or actual molecular spectrum is imported.

Worked examples

Try it yourself

1. Double wavelength while retaining the first-order spacing from the Bragg example. Find the new reported 2θ.

   Show the answer and reasoning

   About 86.32036°.

   sinθ_new=2 sin20°; solve the inverse sine and double the angle. The reported angle need not simply double.

2. Double d in the two-point model. How does the first positive zero expressed in reciprocal coordinate q change?

   Show the answer and reasoning

   Its q value is halved.

   The first zero satisfies qd=π, so q_new(2d)=π.

3. Represent the entire image at twice the pixel count with consistent calibration and uncertainty transformation. What are the new pitch and feature count, and does L or its represented standard uncertainty change?

   Show the answer and reasoning

   Pitch 0.200 nm per pixel and feature count 250; L stays 50.0 nm and u_L stays 1.28062485 nm.

   Pitch and its uncertainty halve, while feature count and its count uncertainty double. Both length and the two propagated uncertainty contributions remain invariant.

4. Remove the verified proton-equivalence constraint from the two-candidate exercise. Which candidates remain?

   Show the answer and reasoning

   Both propanal and propanone.

   Both meet the formula and carbonyl constraints. Missing independent evidence cannot be supplied by restating those same observations.

What to carry forward

• Diffraction, microscopy, composition and spectra constrain different features.
• Declare wavelength, diffraction order and θ versus 2θ.
• Intensity need not retain every phase distinction.
• Pixel pitch is distinct from resolution and complete uncertainty.
• A restricted candidate inference should name its universe and assumptions.