CVD vs HPHT: How the Two Lab-Grown Diamond Methods Differ

HPHT presses are built to contain pressures equivalent to those at depths of around 150 kilometres in the Earth's mantle. Three press geometries are in commercial use¹.

The cubic press uses six hydraulic anvils arranged on the faces of a cube. A small cell containing the carbon source, metal catalyst, and seed sits at the centre. Pressure is applied symmetrically from all six anvils. The belt press uses a pair of large opposing tungsten-carbide anvils with a cylindrical die, the older General Electric design and still in use for industrial diamond synthesis. The BARS press, a Russian split-sphere design, achieves higher pressures in a smaller cell and is the geometry behind much of the modern Russian and Chinese gem-grade HPHT production.

CVD reactors are vacuum chambers, not presses. The chamber holds a heated stage on which the diamond seed plate is mounted. Three power configurations are used¹: microwave plasma CVD (MPCVD), in which a 2.45 GHz or 915 MHz microwave field excites a glowing plasma above the seed; hot filament CVD, in which a tungsten or tantalum filament heated to about 2,200 degrees Celsius dissociates the gas; and DC arc plasma jet CVD. MPCVD dominates current gem-grade production because it produces the cleanest plasma without contaminating the chamber.

The two methods are chemically dissimilar at the atomic scale¹.

In HPHT, the carbon source dissolves into a molten metal flux at high temperature and pressure. Diffusion through the flux carries carbon atoms to the seed surface, where they precipitate onto the growing crystal face. Because growth is from a liquid melt with a roughly isotropic environment around the seed, the resulting crystal habit tends toward the equilibrium octahedral or cuboctahedral form, with relatively flat large facets. Some metal flux is invariably incorporated into the crystal, often visible under magnification as small dark inclusions.

In CVD, the chemistry is gas-phase, dominated by atomic hydrogen². Methane decomposes in the plasma into reactive carbon-bearing fragments, primarily methyl radicals. These deposit on the diamond surface, while atomic hydrogen simultaneously caps dangling bonds and preferentially etches non-diamond carbon faster than diamond, suppressing graphitic growth. The chamber chemistry is therefore self-cleaning at the growth front.

The result is layer-by-layer epitaxial growth on a flat surface, propagating upward at micrometres per hour. The crystal habit is dictated by the seed plate, not by equilibrium thermodynamics, which is why CVD-grown crystals come out as roughly cuboid plates rather than octahedra. Microscopic phase-boundary growth lines run parallel to the original seed surface, giving CVD-grown rough its characteristic layered fingerprint.

The dominant impurity in diamond is nitrogen, and its presence and configuration determine type classification⁴. CVD's gas-phase, tightly controlled chemistry naturally excludes nitrogen unless deliberately introduced. The default product is Type IIa, the same nitrogen-free configuration that the rarest natural diamonds (Cullinan, Lesedi La Rona) belong to.

HPHT's catalyst-flux chemistry sits in a different regime. Iron, nickel, and cobalt all have non-trivial nitrogen solubility, and the press cell almost always contains residual atmospheric nitrogen. The result is that HPHT-grown rough is often Type Ib (isolated nitrogen impurities), giving the stone a yellowish tint at higher concentrations. Boron-getter additions or deliberate boron addition can produce Type IIb (boron-doped, sometimes blue), and rigorous nitrogen exclusion can yield colourless Type IIa.

From a buyer's perspective, the practical implication is straightforward: most modern colourless lab-grown diamonds at higher carat weights are CVD-grown, while HPHT continues to dominate the small-stone melee market where slight tints are less commercially significant.

GIA's gemmological laboratory publications describe characteristic inclusion patterns for each method²³.

HPHT-grown stones often contain small metallic inclusions, the trapped flux. These appear as dark specks under magnification, sometimes elongated, sometimes more rounded. They can be magnetic, and a magnet can in principle pick up a stone with sufficient flux content, although this is a curiosity rather than a routine identification method. Sector zoning between cubic and octahedral growth sectors is sometimes visible.

CVD-grown stones typically show layered growth lines and, less commonly, non-diamond carbon residues if growth conditions are out of optimum. Phosphorus or silicon contamination from chamber components can leave luminescent point defects detectable under photoluminescence spectroscopy. Cathodoluminescence imaging shows the layered growth horizons clearly.

Some CVD-grown rough emerges from the reactor with a brown cast, which post-growth HPHT treatment can remove. The treated stone is no different in chemistry: it is still pure carbon, still in the diamond cubic phase. The treatment alters defect distribution, lifting colour toward colourless².

The terminology trips people up. A diamond can be:

• HPHT-grown: the stone was synthesised by the HPHT method.
• CVD-grown: the stone was synthesised by the CVD method.
• HPHT-treated: the stone, natural or synthetic, was subjected to a post-growth high-pressure-high-temperature process to alter colour.

A stone may be CVD-grown and HPHT-treated, or HPHT-grown and untreated, or natural and HPHT-treated. The grading certificate should disclose every applicable category. FTC rules require treatment disclosures, and consumer-facing language should make the distinction clear (see Chapter 6).

Where this fits in the reference

The next chapter, The 4Cs Explained, sets out how natural and pre-October-2024 lab-grown diamonds are graded under the GIA system. Chapter 4 compares the two laboratories that handle most of the world's diamond grading, and Chapter 5 explains why GIA stopped using the 4Cs scale for new lab-grown reports in October 2024.