Evaporation and sputtering both belong to physical vapor deposition (PVD). Both transfer metal from a source material onto the surface of a sample. However, "being able to form a coating" does not mean the resulting films are identical.
The differences can be examined by addressing three key questions: How does material leave the source? How do particles travel across the vacuum chamber? After reaching the surface, how much of their energy and directional information is retained?
01｜Do Not Rush to Compare Merits and Demerits - First Understand How Metals Are Generated
Thermal evaporation relies on resistively‑heated boats, filaments or crucibles to heat the source material until it develops sufficient vapor pressure.
Electron‑beam evaporation uses a focused electron beam to concentrate energy onto localized regions of the source material, making it suitable for materials that require much higher source temperatures.
Although their heating mechanisms differ, the subsequent fundamental process is similar: the source material transitions into the gas phase, travels through the low‑pressure chamber toward the cooler sample surface, where it adsorbs, nucleates, and gradually coalesces into a thin‑film layer.
Magnetron sputtering does not "boil" the target material. The working gas is excited to form plasma. Ions gain energy near the target surface and bombard the target. Collision cascades knock target atoms out from the surface. The ejected particles then travel through the working gas and reach the sample. The primary function of the magnetic field is to confine electrons more effectively near the target surface, boosting ionization and discharge efficiency.
The actual flux deposited onto the sample is still affected by target geometry, magnetic‑field distribution, working pressure, target‑to‑substrate distance, fixture position, and sample motion mode.
02｜Directionality Is Not a Device Specification, but a Consequence of Particle Transport
In conventional high‑vacuum evaporation, gas‑phase collisions between the source and the sample are rare. Particles travel nearly in straight lines. Surfaces of the sample facing the source readily receive particle flux, whereas areas shadowed by steps or photoresist sidewalls may experience shadowing effects. This is the well‑known "line‑of‑sight deposition".
However, directionality is not determined merely by the term "evaporation". Source dimensions, source‑to‑sample distance, sample tilt angle, fixture shadowing, and rotation mode all alter the incident angle of arriving particles. With tilted substrate placement or planetary motion, the same evaporation system can actively improve side‑wall coverage for certain features.
Sputtering requires sustained electrical discharge within a specific working‑gas environment. After target atoms leave the target surface, they may collide with gas atoms and have their trajectories redistributed. Furthermore, the target itself is not an ideal point source. Consequently, conventional sputtering typically delivers a broader angular distribution of incident particles to the sample compared with standard evaporation. Nevertheless, the statement that "sputtering has no directionality" is equally inaccurate. Lowering or raising the working pressure, adjusting the target‑to‑substrate distance, employing collimating structures, and repositioning the sample all modify collision frequency and incident‑angle distribution. Higher pressure can broaden particle angular spread and may also cause arriving particles to lose more energy.
03｜Why One Process Favors Lift‑Off While the Other Performs Better for Step Coverage
Differences in directionality become amplified when encountering three‑dimensional structures. The fundamental principle of lift‑off is to pattern openings in the photoresist layer first, deposit film over the entire substrate, and then dissolve the photoresist to strip away the film on top of the resist together.
For successful lift‑off, the target film on the substrate inside the openings should preferably not form a continuous connection with the film on top of the resist along the sidewalls. Evaporation with strong line‑of‑sight directionality tends to create a thickness break near the resist sidewalls.
04｜What the Substrate Experiences Goes Far Beyond "Simply Being Coated with a Metal Layer"
Evaporation is often loosely described as "gentle", whereas sputtering is regarded as having "strong bombardment".
Both perceptions have factual bases, yet neither holds true independent of equipment and process conditions.
During evaporation, the sample may still experience temperature rise caused by thermal radiation, heat‑source exposure, and deposition duration.
For electron‑beam systems, additional factors such as electrons, X‑rays, source‑material splashing and particles must also be taken into account.
Polymers, photoresists, low‑melting‑point layers or radiation‑sensitive devices should not be deemed safe merely based on the process name. In a sputtering environment, apart from target atoms, fluxes of ions, electrons, reflected neutrals and impurities may also be present. Substrate bias, plasma location and working pressure alter the actual energy input delivered to the sample. Moderate energy can promote surface migration and interfacial bonding, whereas excessive energy may lead to temperature elevation, surface modification, re‑sputtering or damage to sensitive layers.
Therefore, when discussing sample tolerance, one should not merely ask "whether the temperature is high". It is also necessary to consider what the sample surface consists of, whether thin oxide layers or two‑dimensional materials lie underneath, whether the photoresist is still present, whether the device fabrication is complete, and whether post‑process annealing or cleaning is permissible.
05｜Why "Sputtered Films Are Denser" Is Only a Starting‑Point Statement
A thin film does not descend as a solid block of metal. Instead, incoming particles first adsorb and nucleate to form islands. These islands then grow and coalesce, eventually developing into a continuous film. The energy of arriving particles, substrate temperature, deposition rate, incident angle, surface cleanliness, and residual gas all govern the surface mobility of particles, and modify grain size, porosity, roughness, texture and residual stress. The energetic sputtered particles and associated energy fluxes generally lead to expectations of enhanced surface mobility and comparatively dense films.
Nevertheless, sputtered films can also develop porous columnar structures, substantial stress or high roughness under conditions of high working pressure, significant collisional energy loss, unfavorable incident angles, or inadequate surface conditions. Neither are evaporated films inherently porous. Modifications to substrate temperature, deposition rate, surface treatment and auxiliary energy inputs will likewise alter nucleation and growth behavior.
Higher film density does not automatically translate to better‑performing devices. Metallic electrodes may place higher priority on contact resistance and edge integrity; optical structures may focus more on roughness and optical constants; suspended structures are highly sensitive to stress; diffusion‑barrier layers pay close attention to continuity and defect pathways. "Dense" is merely a material description. Ultimately, performance must be evaluated against the functional requirements of the thin‑film application.
Summary
Both evaporation and sputtering can deposit metals, yet they are not interchangeable push‑button options. Evaporation vaporizes materials into the gas phase via heating and exhibits strong line‑of‑sight directionality under typical conditions. Sputtering liberates target atoms through ion bombardment; collisions within the working gas and equipment geometry generally yield a broader range of arrival angles.
Claims such as "evaporation is better" or "sputtering is better" are premature. A suitable process route should explain how particles arrive at the features, specify which location will be inspected for minimum film thickness, what types of failures to anticipate, and what final criteria will be used to judge whether the thin‑film layer has truly fulfilled its intended function.
