The Five Things That Make a Heat Exchanger Better, and What Makes Them Challenging
Fundamentally, a heat exchanger performs better when heat crosses a thinner, more conductive wall, spread over more surface, through clean channels that share flow evenly. Sounds simple enough, but these goals often overlap and even compete. A thin wall still needs to hold high pressure, and a denser array needs more pumping power to push fluid through it. No heat exchanger finds the perfect equilibrium between the “Five,” since trade-offs exist in every design. The true needle-mover is often the manufacturing process behind the design.
In this article, we break down the “Five” aspects that matter most for heat transfer performance: why they matter, why they’re hard to manufacture conventionally, and what pulsed electrochemical machining (PECM) can do to help.
Key Takeaways
Wall conduction resistance scales with thickness divided by thermal conductivity, so thinner walls pay off most in low-conductivity alloys like Inconel and titanium.
Compact heat exchangers are usually defined by a surface-area density above roughly 700m²/m³, which pushes designs toward small channels that are hard to make and harder to inspect.
Material selection is a negotiation between conductivity and strength at temperature, corrosion resistance and cost.
Burrs, recast and roughness inside channels change heat transfer, pressure drop, fouling and fatigue behavior, even though they rarely appear on the drawing.
In parallel-channel cores, consistency between channels matters as much as the quality of any single channel.
1. Wall Thickness: The Shortest Path Wins
Why it matters
The first of the “Five” is wall thickness. Within a heat exchanger, the thinner the wall between two fluids, the less it resists heat flow. Conduction resistance through a flat wall equals its thickness divided by the material's thermal conductivity (per unit area), so halving the wall predictably halves that resistance.
How much this matters depends heavily on the material. Think of heat crossing an exchanger as passing through three barriers in a row: the hot fluid's boundary layer, the metal wall and the cold fluid's boundary layer. The wall only matters if it's a large barrier compared with the two fluid layers. In a typical liquid-to-liquid exchanger, each fluid layer contributes about 0.0002m²·K/W of resistance (assuming film coefficients around 5,000W/m²·K). A 1mm copper wall (roughly 390W/m·K) adds about 0.000003m²·K/W, which is close to nothing. The same 1mm wall in Inconel 718, at about 11W/m·K according to the Special Metals technical bulletin, adds about 0.00009m²·K/W, nearly half the resistance of an entire fluid layer.
Why should an engineer care? Because it tells you where wall thickness is worth fighting for. In a copper exchanger, you can leave the wall thick and lose almost nothing. In Inconel, the wall is one of the biggest barriers to heat flow, so every fraction of a millimeter you remove shows up as performance.
Put plainly: thinning the Inconel wall from 1mm to 0.5mm raises the overall heat transfer coefficient by roughly 10%, while doing the same to the copper wall changes almost nothing. In the high-temperature, corrosion-resistant alloys that aerospace and energy hardware depend on (the key focus of this article and of PECM's applicability), wall thickness is one of the largest single terms in the thermal budget.
In a copper exchanger, the wall is almost free. In Inconel, it's one of the biggest line items in the thermal budget.
Thinner walls also mean less mass and smaller cores, which counts for a lot in flight and space hardware. On an aircraft, every kilogram of heat exchanger is carried for the life of the airframe, so extra mass turns into extra fuel burn on every flight. Size matters just as much, because engine nacelles, environmental control bays and satellite buses leave very little room for thermal hardware. A thinner wall lets designers hit the same heat duty with a smaller, lighter core, or fit more heat transfer into the envelope they already have. For launch vehicles and spacecraft, where each kilogram adds to launch cost, that trade is often what decides whether a design is viable at all.
Why it's hard to manufacture
Thin walls are notoriously challenging for most, if not all, conventional material removal processes. Cutting tools (and any other contact-based method) push on the part, and a very thin wall deflects and chatters under that load, which makes the final thickness hard to hold. Cutting also leaves residual stress near the surface, and when the part comes out of its fixture, that stress can relax and distort the wall.
EDM and similar methods avoid cutting force, but they leave a recast layer: molten material that re-solidifies on the part surface, creating a brittle, rough exterior. EDM and laser processing can also add a heat-affected zone (HAZ), where the material's microstructure is altered and becomes more prone to microcracking.
These issues compound in AM-printed heat exchangers, whose as-printed roughness and porosity take up more of a thin section than a thick one.
Where PECM fits
PECM removes metal by controlled anodic dissolution with no contact between the tool and the part. With no cutting force, nothing pushes a thin wall out of position during machining, and with no heat input, no heat-affected zone or recast layer forms. Voxel applies this to narrow channels and thin-wall features in alloys like Inconel. Pressure containment and fatigue design still set how thin a wall can safely be, so PECM removes a manufacturing limit while the design limits stay in place.
2. Surface-Area Density: More Transfer Surface per Unit Volume
Why it matters
Heat transfer scales with surface area. Rather than making each heat exchanger extremely large, packing more surface into the same volume is the most direct route to an efficient exchanger that is also light and small. Engineers measure this as surface-area density, in m²/m³. The common benchmark comes from R.K. Shah, who defined a compact heat transfer surface as one above 700m²/m³, as cited in the Wiley book Compact Heat Exchangers. Conventional shell-and-tube units sit well below that benchmark, and this article focuses on compact heat exchangers.
So what governs density? Channel size is the key driver. For a core built from parallel channels, surface-area density is roughly four times the open fraction divided by the hydraulic diameter. In other words, a core with 1mm channels and half its volume open lands near 2,000m²/m³. Shrink the channels to 0.5mm and it doubles to about 4,000m²/m³. Offset strip fins, pin-fin arrays and AM lattices such as gyroids reach high densities through different geometry, yet the principle holds for all of them: smaller features give more surface.
The push for density is why several types of compact heat exchanger exist:
Printed-circuit heat exchangers (PCHEs) are stacks of thin metal plates with fine channels etched into them, bonded into one solid block. They handle very high pressures and show up in LNG, hydrogen fueling and supercritical CO₂ power systems.
Microchannel cold plates are metal plates with sub-millimeter liquid channels that sit directly on hot electronics, such as the processors in AI data centers, and carry heat away into a coolant loop.
Compact recuperators recover heat from a turbine's exhaust and use it to preheat the incoming air, which raises the efficiency of microturbines and some aircraft and power-generation engines.
AM lattice cores are 3D-printed cores built from repeating structures like gyroids, which interleave two fluids in shapes that can't be machined or assembled from plates.
Density is also what makes these designs so demanding to machine and maintain. Smaller channels block more easily and respond more strongly to small dimensional errors, and most of them can't be inspected by eye at all.
Unlike a shell-and-tube unit, a compact core keeps most of its heat transfer surface in passages no one will ever see directly.
Why it's hard to manufacture
Dense features multiply every per-feature cost in the manufacturing process. Microdrills break more often as diameters shrink, and cycle time climbs with feature count when each channel is cut one at a time. EDM electrodes erode as they work, so channel #300 may not match channel #1 unless the electrode is replaced. Printed-circuit heat exchangers take another route: their channels are photochemically etched into thin flat plates, which are then diffusion-bonded into a solid block. Chemical Processing describes channels as fine as 25µm made this way. The method works well, although wet etching cuts sideways as it cuts down, which pushes channels toward shallow, rounded profiles whose depth is bounded by the plate.
Where PECM fits
PECM forms many features in a single pass. In heat exchanger work, one custom cathode machines tens or even hundreds of slots at a time, and with minimal tool wear, a single tool produces thousands of identical parts (more in our article Tool Wear, Repeatability, and the Scalability of PECM). Unlike etching, PECM machines into solid stock and contoured surfaces, so channel depth isn't capped by sheet thickness. It's worth noting that PECM is often not cost-competitive with parts that are already photochemically etched successfully. PECM becomes beneficial when features are too dense or too deep for PCE.
Section digest · Thinner walls give the biggest gains in low-conductivity alloys. · Compact cores start around 700m²/m³, and density climbs as channels shrink. ·Contact-based and thermal processes both get harder as walls thin and feature counts rise. |
3. Material: Conductivity Versus Everything Else
Why it matters
Predictably, the choice of material is incredibly important for a heat exchanger. Thermal conductivity sets how freely heat moves through the wall, but surprisingly, it's rarely the property that decides which material gets used. Copper and aluminum are the best conductors among common engineering metals, but both lose strength quickly as temperature rises, which rules them out of many heat exchanger applications. Stainless steels, titanium alloys and nickel superalloys (e.g., Inconel) conduct heat poorly, yet they hold their strength and resist oxidation and corrosion in conditions copper can't survive. Refractory metals like tungsten and molybdenum are sometimes good choices, pairing reasonable conductivity with extreme melting points, though they oxidize in hot air and are brittle at room temperature (see Plansee's tungsten property data).
The real takeaway? Material and wall design can't be chosen separately. Specifying Inconel for a 650°C gas stream makes the wall-thickness lever we discussed earlier far more important, because the material itself is a poor conductor. Specifying copper for a liquid cold plate, on the other hand, moves the bottleneck to surface area and flow.
Material | Thermal conductivity (W/m·K, approx., room temp.) | Strength at temperature | Typical heat exchanger roles | Machining challenges |
Copper (C101/C110) | ~390 | Low; softens above roughly 200°C | Electronics cold plates; copper alloys in rocket chamber liners | Gummy, smears, burrs easily |
Aluminum 6061 | ~167 | Low above roughly 150°C | Automotive, aircraft environmental systems, brazed plate-fin | Burrs, built-up edge |
316L stainless | ~16 | Moderate to good | Process, semiconductor, pharma, printed-circuit HX | Work hardening, burrs |
Ti-6Al-4V | ~7 | Good to roughly 400°C | Aerospace, seawater, chemical service | Cutting heat concentrates at the tool; galling |
Inconel 718 | ~11 | High, to roughly 700°C | Turbine hardware, recuperators, AM heat exchangers | Work hardening, rapid tool wear |
Tungsten / molybdenum | ~170 / ~140 | Very high in vacuum or inert gas | Specialty high-temperature and vacuum hardware | Brittle; edge chipping and cracking |
Why it's hard to manufacture
Every material family brings its own machining problem. Inconel and titanium work-harden and wear through cutting tools, and titanium's low conductivity traps cutting heat right at the tool edge (our deep dive on Inconel covers this in detail). Copper and aluminum are soft and ductile, so they smear and form burrs rather than shearing cleanly. Tungsten and molybdenum are hard and brittle at room temperature, and their edges tend to chip or crack, as we described in All About Refractory Metals.
Those problems get worse once the features are thin or dense. In Inconel, the heat a cutting tool generates stays at the cutting edge instead of leaving with the chip, so tools wear quickly and have to run at low speeds (American Machinist covers why). Each pass also hardens the surface the next pass has to cut. On a thin wall, that means fighting deflection and work hardening at the same time, often with light cuts and slow feeds that stretch cycle time. On a dense array, every additional slot or hole adds to tool consumption, and the tool wears through the run, so dimensions drift between the first feature and the last.
Refractory metals bring a different risk. Tungsten has a ductile-to-brittle transition temperature well above room temperature, so it behaves like a ceramic under a cutting tool, and thin walls or closely spaced features are prone to chipping and cracking (Industrial Metal Service). EDM is often the fallback, but it's slow on fine features and leaves recast on a material that's already crack-sensitive.
At production scale, those effects multiply. Tool changes, re-measurement and in-process inspection all add labor to every part. A crack or out-of-tolerance wall found late in the process scraps a part that already carries the cost of expensive material and every earlier operation. For a single prototype, shops can work around these problems with patience. For thousands of identical heat exchanger cores, they become the deciding cost.
Where PECM fits
This is arguably where PECM's advantage is clearest, because PECM's ability to remove a material isn't tied to its hardness. Electrochemical dissolution is driven by the material's electrochemistry, so a work-hardened Inconel surface machines on the same terms as an annealed one. A 1.5mm thin wall in Inconel 625 is machined at roughly the same rate as the same feature in copper.
The requirement, of course, is electrical conductivity: PECM works on conductive metals, which leaves ceramic and polymer exchangers outside its scope. Microstructure still shapes results, and What Truly Determines Whether PECM Handles an Alloy Well? explains why. Voxel has regularly produced 300µm microholes in copper and 0.125mm-thin walls in stainless steel.
4. Surface Condition: The Hidden Driver
Why it matters
A heat exchanger channel's surface condition changes how well it moves heat and how much pressure it costs. Surface roughness disturbs the boundary layer, which can raise heat transfer and always raises friction. A DOE study compared a smooth drawn tube against additively manufactured channels for supercritical CO₂ heat exchangers. The AM channels had roughness heights equal to 0.29–0.73% of their hydraulic diameter, which works out to roughly 3–7µm of roughness for every millimeter of channel width. The rough channels showed friction factors about 8% above smooth-pipe correlations, and standard heat transfer correlations missed the measured values by roughly 25–30%. The roughest channel still delivered about a 10% gain in heat duty over the smooth tube. Put plainly: even relatively minor roughness differences have real effects on a heat exchanger, and they're notoriously difficult to predict and control.
That was roughness on a minute scale. On a larger scale, burrs and debris cause a different kind of damage. A burr at a channel inlet narrows the opening and steers flow away from that channel, while loose particles and partially fused powder can break free in service and lodge downstream. Rough or contaminated surfaces also give fouling deposits a place to start. Fouling is the gradual buildup of scale, corrosion products, particulates or biological growth on heat transfer surfaces, and it acts like an insulating blanket that drags performance down and raises pressure drop over time. It's also expensive: a widely cited estimate from Müller-Steinhagen and colleagues, based on studies from the 1980s and 1990s, put its cost at roughly 0.25% of GDP in industrialized countries (Reliable Media). All of this ties back to the recast layers and heat-affected zones we discussed earlier, which add a durability problem, since their microcracks become crack-start points under thermal cycling (see Microcracks: Their Scale & Importance).
Why it's hard to manufacture
All surface finishing can be difficult for manufacturers, but the worst cases are internal, non-line-of-sight areas. A burr or other irregularity where two drilled passages intersect sits out of reach of brushes and hand tools. As-printed AM channels carry several micrometers of Ra roughness and loosely attached powder along their full length. Post-processing methods aren't always reliable either. Abrasive flow machining smooths channels but can leave media behind and round edges unevenly. Electropolishing removes material well, but it does so from every exposed surface, which shifts dimensions on features that were already at size, since the process can't be selective.
Where PECM fits
PECM excels at creating burr-free features with no recast layer because of how it works: it removes workpiece material atom by atom, producing a mirror-like surface finish as part of the process, with no extra steps needed. Voxel has regularly produced surfaces on tough materials down to 0.005–0.4µm Ra and repeats this in challenging internal features at high volumes.
For additively manufactured parts with non-line-of-sight finishing needs, Voxel's oscillatory PECM (OPECM) finishes enclosed internal channels. On Inconel 718 serpentine channels, it brought wall roughness from about 5µm Ra down to roughly 0.24–0.45µm Ra (Enclosed, Internal AM Finishing & Voxel's OPECM Methodology).
Have a heat exchanger or cold plate with a channel-quality problem? Contact us at info@voxelinnovations.com.
5. Uniformity: Influencing Flow and Pressure
Why it matters
Every gain from the first four levers is paid for in pressure drop. Smaller channels and more surface raise friction, and more friction means more pump or compressor power taken straight out of overall system efficiency. A good design finds the balance point where extra heat transfer stops being worth the extra pressure loss.
Uniformity matters just as much. Parallel channels share an inlet header and an outlet header, and flow divides among them according to each channel's resistance. A channel that's slightly narrower or partly blocked receives less flow than its neighbors, which leaves part of the core underused and can create local hot spots (engineers call this flow maldistribution). In narrow laminar channels, the cost is steep: at a fixed pressure difference, flow through a round channel scales with roughly the fourth power of its diameter, so a channel 5% smaller than its neighbors carries about 19% less flow.
Why it's hard to manufacture
Holding every channel to the same geometry, understandably, is hard at production scale. Drill wear shifts diameters gradually across a run, and fixturing these components can be challenging depending on their size or material. Laser-drilled holes often taper, and their entrances vary from hole to hole. EDM electrodes erode, so later features drift from early ones. PCE does a decent job of maintaining uniformity, but it's limited to a certain type of geometry, so it doesn't apply to many designs.
Where PECM fits
With minimal tool wear, PECM holds geometry steady over long production runs, and features machined in the same pass share the same process conditions. That consistency is what turns a uniform design into a uniform part, a topic we covered in a recent article, The Role of Uniformity in Critical Manufacturing. Repeatability depends on more than the cathode, though. Electrolyte flow and fixturing matter as well, and both are part of every fit review.
Section digest ·Roughness and burrs change heat transfer and friction in ways standard correlations don't capture well. ·Fouling and maldistribution both start at small surface and dimensional defects. ·In parallel laminar channels, a channel 5% smaller can carry roughly 19% less flow. |
Every Lever Depends on the Part Being Made Right
Manufacturing limits every one of the “Five.” A design calling for thin walls, thousands of uniform channels and clean internal surfaces only performs as modeled if production delivers that geometry on every part.
A strong apparent PECM fit still needs a look at the specific part: geometry, cathode and tooling strategy, electrolyte flow, tolerance stack, inspection needs and expected volume. A borderline application isn't an automatic no either. Those are the variables our team works through on every new heat exchanger and cold plate we review.
Capability Snapshot: PECM for Heat Exchanger Features
Capability | What Voxel has published |
Parallel features | Tens to hundreds of slots per custom cathode, per pass |
Tool life | Thousands of identical parts from one cathode (minimal tool wear) |
Thin walls | 0.125mm walls in stainless steel |
Surface finish | Down to 0.005–0.4µm Ra on tough materials |
Internal AM channel finishing (Inconel 718, OPECM) | About 5µm Ra reduced to 0.24–0.45µm Ra |
Microchannels | 1mm × 2.5mm channels in Inconel 625 |
Microholes | 300µm microholes in copper |
Thermal effects | No heat-affected zone, no recast layer |
Edge condition | Burr-free, with a specifiable edge radius |
Materials | Conductive metals, including stainless steel, titanium, Inconel, copper, tungsten and molybdenum |
Figures come from Voxel's published work. Results depend on the specific geometry and material.
Send Us the Channel That's Limiting Your Design
Working on a heat exchanger, cold plate or manifold where wall thickness, channel density or internal finish is holding the design back? Send Mike and Kirk the material, channel geometry, wall target, current process and expected volume, and we'll tell you whether PECM is worth a closer look. You can also request a 15-minute fit discussion at info@voxelinnovations.com.
Comments