CAYOM CA-4 vs. BOLIGO Z980 Thermal Pastes Review – More than just similar bargains?

1 - Introduction and overview 2 - Shear behavior, microscopy, and chemical composition 3 - Lab measurements, interface resistance and bulk thermal conductivity 4 - Real performance on CPUs, notebook chips, and GPUs 5 - Durability (pump-out), summary, and conclusion

Application pattern and pressing behavior of both pastes

Next, I evaluate the visible application pattern and the pressing behavior. Both pastes were applied to the test specimen under identical laboratory conditions and then pressed, so that the differences can be derived directly from the deformation, the squeeze-out behavior and the remaining layer structure. The initial application of CA-4 reveals a relatively homogeneous, soft and finely structured surface. The paste flows evenly over the top of the stamp and fills the surface without any recognizable cavities. This homogeneity is consistent with the previously observed fine particle morphology and the narrow size distribution of the fillers. Even in the initial state, it can be seen that the binder phase is sufficiently smooth to embed the particles evenly. Even in the initial state, however, the Z980 shows a much more uneven application pattern. Although the paste covers the surface, it appears more grainy and less smooth. The surface shows clear traces of the spatula movement and appears less homogeneously mixed. This visual unevenness corresponds exactly to the previously documented agglomerated particle structures in the microscopy.

During pressing, the CA-4 is distributed laterally in a controlled manner and without abrupt break-out. The lateral squeezing is symmetrical and forms a round, evenly sloping bead. The edges do not tear in an uncontrolled manner, but remain structurally closed. The surface of the remaining layer appears smoothed, but continues to show fine, linear flow patterns along the direction of pressing. These patterns are airy and flat, indicating a moderate viscosity that offers the user a wide tolerance range. It is clear from the pressing behavior that the CA-4 is comparatively user-friendly. It forgives uneven pressure, spreads reliably over the entire surface and is not prone to local thinning or splitting. For inexperienced users, this means there is a low risk of creating voids or air pockets. The CA-4 is therefore relatively resistant to typical application errors.

When pressing the Z980, there are characteristic differences to the CA-4. The laterally pressed bead appears more uneven, inhomogeneous and partly segmented. Local constrictions and asymmetrical breakouts occur. The bead tears in places, which indicates lower internal cohesion and a less elastic matrix. The flow movement is not uniform across the entire width, so that edge areas are compressed more than the central section. The risk of local material displacement is visibly higher. The user must therefore dose this paste more carefully and apply it more neatly to avoid areas that are too thick or too thin. The Z980 requires more precise control of the application pressure and quantity. Without this control, uneven layer formation or partial thinning can occur during assembly, which in turn affects thermal performance.

When applied and under pressure, CA-4 presents itself as a mature, free-flowing paste that spreads evenly, even for less experienced users. Its homogeneity leads to stable layer structures and low error tolerance during application. Z980, on the other hand, exhibits a behavior that calls for caution. The coarser particle structure, the greater variation of the matrix and the more uneven pressing pattern require more experience and attention. A user who does not reliably control the layer thickness or does not apply the pressure evenly risks uneven layer profiles and possible variations in performance. The processing of both pastes thus shows a consistent picture with all previous findings: finer dispersed and more homogeneous pastes such as CA-4 can be processed with greater technical precision, stability and user-friendliness, while coarser and more heterogeneous systems such as Z980 require greater care and experience during application.

Why I compare the two pastes with the Arctic MX-6

I deliberately compare the CAYOM CA-4 and the Boligo Z980 with an established reference paste such as the Arctic MX-6, as this results in several analytical advantages. In terms of price and technology, the MX-6 is in a segment that is relevant for many users who deliberately do not opt for premium products with proven industry certifications. This creates a sensible frame of reference that has an impact on suitability for everyday use as well as on processing and thermal behavior.

The MX-6 is also a well-documented product with clearly reproducible measurement results. It therefore serves as a consistent reference point for classifying the performance of cheaper alternatives. Both pastes tested, CA-4 and Z980, come from a market environment in which reliable technical data is rare. A comparison with a known quantity makes it possible to precisely classify differences and similarities without having to rely on information from marketing sources.

At the same time, the MX-6 is technically a realistic benchmark for modern pastes in this price range. If an inexpensive paste comes close to it or even falls below it, this can be understood as a valid statement about its basic suitability. If, on the other hand, it is visibly above it, the result is clearly related to a product that a wide range of users already know and can assess. The MX-6 therefore acts as a practical reference line that makes the results easier to compare.

The minimum coating thicknesses achieved (BLT) at a glance

The minimum achievable bond line thickness is a key parameter, as a thinner layer with the same material quality reduces the thermal resistance. All three pastes were measured under identical conditions, namely 9 Newtons per square centimeter on a defined 1 square centimeter contact area at a sample temperature of 60 degrees Celsius. At 14 micrometers, the CAYOM CA-4 achieved the lowest layer thickness in the test field. This is consistent with its very uniform, fine-grained structure and good internal cohesion, which enables controlled and far-reaching extrusion without the layer thinning out locally or becoming unstable. The visual structure and the homogeneously distributed particle morphology result in a high capacity for compaction.

The Boligo Z980 is just behind at 15 micrometers. The larger and more unevenly distributed filler slightly limits the compactability, but also prevents visible instabilities. The slightly higher thickness corresponds exactly to the previously observed differences in the particle size distribution and the lower homogeneity. Although the Z980 can be compacted well, it does not achieve the fine structure of the CA-4. At 16 micrometres, the Arctic MX-6 is at the upper end of this comparison. This is a classic value for a well-formulated consumer paste with a medium-fine filler structure. The MX-6 shows a stable, but not maximally compressible layer. Its behavior thus confirms its position as a solid reference paste without coming close to the CA-4 in terms of compressibility. A direct comparison shows that the CA-4 offers the best compressibility, followed by the Z980, while the MX-6 forms slightly thicker layers. The differences are small in absolute terms, but precisely reflect the material structure observed in each case.

The effective thermal resistances Rth_{, eff}

The graph shows the thermal resistance Rth as a function of the layer thickness BLT. This representation is much more meaningful than considering the minimum layer thickness alone, as it shows how efficiently a paste transports heat when it is operated in practical, realistic BLTs. All three pastes follow a linear increase in Rth with increasing layer thickness, which is typical for filled silicone pastes. Each additional micrometer layer lengthens the thermal conduction distance and increases the amount of less conductive binder phase, increasing the overall resistance proportionally. However, it is important to emphasize that this measurement only evaluates the thermal conductive path, not the mechanical stability, pump-out risk or long-term behavior.

The CA-4 is the highest over the entire measurement range. This means that it has the highest thermal resistance with an identical BLT. This is consistent with its elemental composition, which is strongly oxidic and has a very homogeneous, but not maximally conductive particle structure. The red curve is stable and without anomalies, which indicates a well-controlled internal structure, but is at a slight thermal disadvantage. As expected, the MX-6 forms a balanced midfield. The blue curve is visibly below the CA-4, but follows an almost identical gradient profile. This confirms its role as a solid reference paste with a good ratio between filler content, binder phase and thermal performance. Thermally, it is clearly more efficient than the CA-4, but not at the top of this comparison. The Z980 delivers the lowest thermal resistance in this measurement and is therefore consistently below the other two pastes. The green curve marks the best thermal conduction distance with identical BLT. This is remarkable, as its particle structure appears much coarser and more heterogeneous under the microscope. However, the LIBS analysis shows that it has more carbon-containing binder components in its composition as well as an altered balance between aluminum, oxygen and zinc. This combination can lead to locally larger particle contact surfaces, which reduce the thermal resistance in the pure ASTM measurement setup.

The Z980 shows the lowest values in this pure Rth comparison, followed by the MX-6. The CA-4 is just above it and is thermally the weakest of the three pastes, even if the differences in the practically relevant range remain moderate. These results are consistent with the previously determined structural properties. The CA-4 is extremely homogeneous, but thermally conservative. The Z980 is mechanically and structurally more unstable, but thermally surprisingly efficient and the MX-6 forms the expected, stable mean value.

I have now compared the relevant layer thicknesses from 25 to 400 µm as a bar chart for _Rth.

Evaluation of the effective thermal conductivity λ_eff

The graph of effective thermal conductivity λ_eff shows how efficiently the three pastes transport heat when the entire real thermal conduction path is taken into account. Unlike the nominal thermal conductivity from the manufacturer’s specifications, λ_eff shows exactly what is relevant in practical use, including particle contact, binder content and compaction properties.

The Arctic MX-6 achieves the highest effective thermal conductivity across all measured BLTs. Even with thin layers, it is visibly above the other two pastes and extends this lead further with increasing BLT. This indicates very efficient particle cross-linking, which remains stable even when the layer becomes thicker. The MX-6 thus shows the best combination of filler structure and binder content in terms of pure heat flow. Its progression is extremely harmonious and increases evenly, which speaks for a well-calibrated formulation.

The Z980 is in the middle of the field, but much closer to the MX-6 than to the CA-4. Its λ_eff increases rapidly with increasing BLT and without any noticeable instabilities. In the lower BLT range, the distance to the MX-6 is even clearer, but is reduced at higher layer thicknesses. This means that the coarser, heterogeneous particle distribution of the Z980 does not necessarily lead to poorer thermal conduction paths. The Z980 therefore offers a thermal performance that can clearly hold its own despite its optical and structural unevenness.

The CA-4 shows the lowest effective thermal conductivity over the entire measuring range. The gap to the other two pastes remains the same across all BLTs. The reason for this lies in the combination of the high oxide filler fraction and the very fine, narrow particle distribution. Although this ensures excellent homogeneity and good mechanical stability, it does not lead to maximum thermal contact surfaces between the particles. The result is a constant but comparatively lower heat transfer rate.

Although the differences remain moderate, they clearly show that a clean, homogeneous structure does not automatically mean maximum thermal performance. Pastes such as the Z980 can clearly benefit from coarser but effectively interconnected particles in certain parameters. Of course, the whole thing is also shown again as a bar chart for the most important layer thicknesses. Just click through and see where the paste is positioned:

Bulk thermal conductivity, interface resistance and quality of measurement

The following is a detailed evaluation of the two regression diagrams for the bulk thermal conductivity of the CAYOM CA-4 and the Boligo Z980. The bulk thermal conductivity results from the slope of the linear regression of the measured thermal resistances over the layer thickness. The CA-4 achieves a value of 4.156 ± 0.067 W/mK, while the Z980 is slightly higher, at 4.304 ± 0.060 W/mK. Although this difference is measurable, it remains within the practically narrow range of two pastes that come from a very similar material class. Both have a clear aluminum silicate filler profile, which explains the comparable conductivity ranges.

The comparison of the regression curves shows that both pastes behave almost ideally linear. The CA-4 has a coefficient of determination of 0.999342816 and the Z980 has 0.999462939. Both values are extremely close to 1 and confirm that the thermal resistances increase in excellent proportion to the layer thickness. Such behavior is characteristic of pastes with a homogeneously distributed particle fraction and without significant structural inhomogeneities along the shear axes. It is only noticeable that the Z980 shows slightly lower scattering, which is possibly due to a slightly more uniform particle distribution or a lower viscosity.

In the practical evaluation, this means that both products deliver a consistent and easily calculable thermal performance. Differences in the range of a few tenths of a watt per meter and Kelvin are hardly relevant in real systems, but they help to classify two products in terms of materials science, which are very clearly attributable to the same manufacturing approach.

It is also interesting to compare the interface resistances of the two pastes. The CA-4 has an interface resistance of 5.2 ± 0.9 mm²K/W, while the Z980 is slightly higher at 6.4 ± 0.7 mm²K/W. This value describes the additional thermal resistance at the interface between paste and substrate, i.e. the zone characterized by roughness, micro-cavities and wetting behaviour. While the bulk value is determined purely by the inner material, the interface resistance determines how effectively the paste comes into contact with the surface and how well it fills surface structures. The lower interface resistance of the CA-4 indicates that it works better into the copper surface during application and leaves less contact loss. This matches the previously observed image impressions of the tear-off structure and the dispersion, where the CA-4 showed a slightly more complete wetting behavior. The Z980, on the other hand, appears with a slightly higher interface loss, which may be due to either the slightly stiffer paste or a less pronounced ability to micro-adapt the surfaces.

In practice, this means that the CA-4 has a slight advantage in initial wetting on rougher or structured surfaces, while the Z980 makes up some ground via its bulk thermal conductivity. In real-life applications, however, both are close to each other and achieve values that are absolutely sufficient for typical consumer cooling systems.