CPK Models, Molecular Shape, and the CineMol 3D Tool
Corey, Pauling, and Koltun gave chemistry space-filling models that changed how scientists see molecular shape. ChemIllusion's CineMol 3D tool brings that same intuition into fast, editable digital teaching visuals — and every render ships with built-in alt text for accessible export to slides and handouts.
In the history of molecular visualization, few tools changed as much as space-filling models did. Before CPK models, chemists had good bonding diagrams and early ball-and-stick kits. What they lacked was an immediate sense of molecular volume — the fact that atoms are not points on a line but spheres that push against each other and fill space.
CPK models fixed that. And their influence is still visible every time a textbook shows a space-filling diagram or a 3D viewer defaults to a sphere representation.
Ibuprofen in CineMol 3D: ball-and-stick and CPK space-filling, both rendered from the same structure at the same orientation. Sticks show you connectivity; spheres at van der Waals radii show you how much room the molecule actually takes up.
Where CPK models came from
The CPK abbreviation stands for Corey, Pauling, and Koltun — three scientists whose work built on each other across two decades.
Linus Pauling developed early physical models at Caltech in the early 1950s to study protein secondary structure. Robert Corey collaborated on that work, and together Corey and Pauling published precise atomic radii for use in physical models. These became the basis for accurate sphere sizes in what we now call CPK coloring.
Walter Koltun formalized and commercialized the approach in the 1960s. He developed a practical plastic model system where each atom type had a defined sphere size (based on van der Waals radii) and a standardized color. Carbon was black. Oxygen was red. Nitrogen was blue. Hydrogen was white and notably smaller. Sulfur was yellow. The system made molecules recognizable by shape and color even before a student read the label.
The sphere sizes in CineMol correspond to published van der Waals radii (Bondi 1964 values are the most widely used reference). The space-filling style renders each atom as a solid sphere at its van der Waals radius, with no gaps between touching atoms — the same logic Koltun encoded in plastic more than fifty years ago.
What space-filling models help students see
A line-angle structure is efficient because it hides detail. Carbon atoms disappear into vertices. Hydrogens are implied. Bond lengths and atom sizes are abstracted away. That abstraction is powerful when the goal is mechanism or connectivity — but it can obscure the spatial reality that drives many organic chemistry ideas.
Space-filling models bring the hidden parts forward. They make steric crowding visible. They show why two groups cannot occupy the same region of space. They help explain why a bulky nucleophile approaches one face of a molecule more easily, why conformations differ in energy, and why a flat aromatic drawing is only part of the story.
For dimethylformamide, a flat structure hides the partial double-bond character of the C–N bond and the restricted rotation that makes DMF a two-signal molecule in NMR at room temperature. A space-filling model makes the nitrogen's geometry visible: the three groups around N lie in a plane, and the lone pair is delocalized into the carbonyl. The amide resonance is not just a mechanism arrow on a page. It is a shape.
CineMol 3D on the ChemIllusion canvas
CineMol is a lightweight 3D renderer built into the Generator. The goal is teaching clarity: take a structure students already have, choose a rendering style, rotate it into a useful orientation, and place the result in the lesson canvas.
The workflow is short:
Draw or paste a molecule — DMF, aspirin, caffeine, whatever the lesson requires.
Open CineMol 3D from the toolbar or by asking the assistant.
Choose a style: Space-filling for CPK-like renders, Ball & stick for bonding clarity, Tube for backbone geometry, or Wireframe for lightweight overlays.
Choose a look: Glossy for polished slides, Cartoon for cleaner lesson figures.
Rotate the model until the relevant shape is visible.
Click Add to canvas — the render is placed as an editable canvas object with alt text metadata automatically set.
CineMol is built on the open-source CineMol library (v1.0.2) by David Meijer at Wageningen University. The library renders small-molecule configurations directly to SVG using a ray-casting-style projection: each atom is drawn as a shaded sphere whose highlight and shadow are computed from a single virtual light source, producing the characteristic glossy look without requiring a GPU or a full 3D engine. Bonds are rendered as cylinders connecting sphere centers, and the entire output is a scalable SVG — crisp at any size, editable in Illustrator or Inkscape, and embeddable in slides without rasterization artifacts.
See also: github.com/aligfellow/xyzrender — xyzrender uses CineMol-style interlocked-silhouette vdW sphere rendering for its surface presets.
The render is generated server-side from the molecule's SMILES, so it reflects the actual structure — not a stock image or a hand-drawn illustration. If the lesson is about steric hindrance, the visual emphasizes crowding. If it is about functional groups, the CPK colors identify heteroatoms immediately.
Toggling hydrogens on or off is one click. A structure that hides hydrogens is cleaner for mechanism lessons. A structure that shows explicit hydrogens is more useful when teaching steric arguments or NMR proton environments.
Alt text and WCAG accessibility
When CineMol adds a render to the canvas, ChemIllusion automatically generates descriptive alt text based on the molecule name, rendering style, and color scheme. For a space-filling model of dimethylformamide, the generated alt text reads:
That alt text travels with the object. When the canvas is exported to a slide deck or PowerPoint file, the alt text is embedded in the image shape so screen readers and accessibility checkers can find it. The same applies to Newman projections (see the Newman blog post) and reaction coordinate diagrams (see the reaction coordinate blog post).
Teachers can review and edit the generated alt text before exporting. For a lesson figure where the alt text should not name the molecule (for example, a mystery structure in a quiz), the teacher edits the metadata before sharing. For standard teaching materials, the auto-generated text is a useful starting point and covers the main WCAG requirement: every non-decorative image should have a text description that conveys its purpose.
ChemIllusion is being built with WCAG 2.1 AA alignment and ACR/VPAT readiness. Alt text on generated figures is part of that commitment — not a bolted-on feature, but a default behavior tied to how the tool creates objects.
For teachers: slides, handouts, and LMS
CPK-like renders are useful wherever a 2D structure is doing too much work. A slide can show the skeletal formula beside a space-filling model. A worksheet can ask students to compare which face of a molecule is more accessible. A short GIF can rotate a model just enough to reveal the feature the lesson is about.
Because the render is generated from the molecule rather than assembled as decorative art, it stays tied to the chemistry being taught. That makes it more reusable than a one-off screenshot and faster than building a model in a separate 3D tool.
For LMS integration, CineMol renders export cleanly as PNG images with embedded alt text. When placed in a slide deck through the CourseHouse, the image shape in the resulting PowerPoint file carries the ChemIllusion-generated alt text, making it easier to pass an LMS accessibility check.
Related tools on the same canvas: Newman projections for conformational analysis and reaction coordinate diagrams with real ΔH values. All three tools are accessible from the Generator toolbar and produce objects with built-in accessibility metadata.
Try it
Open the Generator with a molecule already loaded, or draw one. Then open CineMol 3D — either from the toolbar cluster or by asking the assistant. Choose Space-filling style and Glossy look for the classic CPK-inspired render. Rotate until the relevant face or feature is visible, then add to the canvas.
No account required for your first render. The tool is available on all plans.