English
Menu

SADP vs SAQP vs LELE: Pitch Division, Overlay and Mask Count

Multiple patterning forms patterns whose pitch is too fine to resolve in a single exposure by splitting them over several exposures or process steps. LELE splits the layout across two masks and repeats exposure and etch twice so that the lines interleave. SADP builds spacers on the sidewalls of core lines (mandrels), then removes the mandrels, so each mandrel yields two lines and the pitch is halved. SAQP repeats the spacer process twice to divide the pitch by four. LELE gives more layout freedom but demands very tight overlay between its two exposures; SADP and SAQP place lines by self-alignment but produce regular line arrays and need cut masks to remove unwanted parts.

AspectSADPSAQPLELE
Pitch division1/2 (two lines from each mandrel)1/4 (the spacer process is done twice)1/2 (1/3 with three masks, LELELE)
Typical reachable pitch (with ArF immersion)Half the single-exposure limit of about 80 nm, so about 40 nmDown to the low 20s of nanometersAbout 40 nm in principle; in practice limited by overlay
What sets line positionsMandrel position and spacer film thickness (self-aligned)Mandrel position and the thickness of the two spacer films (self-aligned)The position of each of the two exposures; the second is overlaid on the first
Overlay sensitivityLine-to-line positions are not affected by overlay; cut masks must still be aligned to the linesLike SADP, lines are not affected; aligning the cut and block masks is even tighterVery sensitive: any shift of the second exposure directly changes the space to the neighboring line
Line width and space variationLine width is set by film thickness and is uniform; if the mandrel width is off, alternating spaces become unequal (pitch walking)More dimensions contribute (mandrel width and two film thicknesses), so pitch walking is harder to controlThe line widths of the two exposures vary independently, and spaces change with overlay error
Masks (that form the lines)One mandrel mask plus cut or block masksOne mandrel mask plus cut or block masks (usually more than SADP)Two (three for LELELE)
Main process stepsExposure → mandrel etch → spacer deposition (ALD, etc.) → etch-back → mandrel removal → transfer to the underlying layer → cutSADP steps plus a second spacer deposition, etch-back and mandrel removalExposure → etch → exposure → etch (transferred in turn into a hard mask or similar)
Layout freedomLow: basically regular lines in one direction, with unwanted parts removed by cutsEven lower than SADP: limited to nearly uniform gratingsHigh: can handle 2D shapes, but arrangements that cannot be split into two colors (such as three mutually close patterns) are not allowed
Typical usesFins, gates, metal lines and fine lines in memoryThe finest fins, the densest metal layers and memory linesContacts, vias and metal layers with complex shapes

Figures are typical orders of magnitude and vary by process generation and product.

When to use which

SADP
When you need regular lines in one direction at half the pitch with uniform line width. Dense lines can be made without worrying about overlay between them.
SAQP
When even SADP is not enough and the finest line pitch is needed. It has the most steps and cost, but reaches pitches in the low 20s of nanometers even with ArF immersion.
LELE
When you need to split 2D patterns that are not regular lines, such as contacts and vias, across exposures. Overlay must be controlled very tightly.
Relation to EUV
EUV lithography brought many layers back to a single exposure and reduced multiple patterning steps. At the very finest pitches, EUV is still sometimes combined with spacer techniques.

FAQ

Why is SADP robust to overlay error?

Because each pair of lines comes from spacers formed on both sidewalls of the same mandrel. Their positions are set by the mandrel position and the spacer thickness, with no second exposure to align. In LELE, any shift of the second exposure directly changes the space to the neighboring line.

What is pitch walking?

It is when alternating spaces become unequal in spacer-based patterning. The space left where a mandrel was removed is set by the mandrel width, while the space between neighboring mandrels depends on the mandrel pitch, the mandrel width and the spacer thickness, so if the mandrel width misses its target the two kinds of space differ. SAQP has more contributing dimensions, so it is even harder to control.

Why are cut masks needed?

In SADP and SAQP the spacers wrap around the ends of the mandrels and form loops, leaving only a regular array of lines. To get the shapes a circuit needs, line ends and unwanted parts are cut or blocked with another mask. That mask is placed relative to the lines, so it needs good overlay.

What arrangements cannot be split into two colors in LELE?

Arrangements that form an odd loop, such as three patterns that are each too close to the other two for a single exposure. However they are assigned to two masks, two patterns that are too close end up on the same mask. Design tools flag such arrangements as errors to be fixed.

Check your understanding

Which correctly describes double patterning (multiple patterning)?

  1. ATwo wafers are exposed at once
  2. BPatterns with a pitch too fine to resolve in one exposure are formed in multiple exposure and processing steps
  3. CCoating two layers of resist doubles the resolution
Show answer and explanation

Answer: B. Patterns with a pitch too fine to resolve in one exposure are formed in multiple exposure and processing steps

Double patterning forms pitches too fine for a single exposure either by splitting the layout into two interleaved patterns exposed and etched separately (LELE) or by doubling lines with sidewall spacers (SADP, SAQP). It pushed ArF immersion beyond its roughly 80 nm pitch limit into the 20 nm-plus range, but at a steep increase in steps, masks and cost, driving the adoption of EUV.

In LELE (litho-etch-litho-etch) double patterning, what is especially strictly required?

  1. AResist color
  2. BThe number of exposure tools
  3. COverlay accuracy between the first and second patterns
Show answer and explanation

Answer: C. Overlay accuracy between the first and second patterns

LELE splits a dense pattern into two separately exposed and etched patterns, so any misalignment of the second relative to the first directly changes the spacing between neighboring lines, affecting wire capacitance and shorts—a 2 nm shift on a 20 nm space is a 10% change. Extremely tight overlay is therefore required. Spacer-based SADP self-aligns line positions and largely avoids this problem.

Which technique, used to make the thin fins of FinFETs, uses sidewall spacers to create patterns finer than the lithography limit?

  1. ASADP/SAQP (self-aligned double/quadruple patterning)
  2. BOPC
  3. CCMP
Show answer and explanation

Answer: A. SADP/SAQP (self-aligned double/quadruple patterning)

SADP coats a mandrel pattern with a conformal film (e.g., by ALD), forms sidewall spacers, then removes the mandrel, doubling line density. SAQP repeats this to quadruple it. Because line width is set by film thickness rather than exposure, variation is small, and fins and wires finer than ArF immersion's limit (about 80 nm pitch) become possible. Even with EUV, these techniques are combined for the tightest pitches.

  1. For layouts intended for double patterning, which correctly describes how design tools assign patterns?
  2. Based on the resolution formula R = k1 × λ / NA, how can resolution be improved (R made smaller)?

Practice Lithography questions

Key terms explained

Multiple patterning
Forming patterns with a pitch too fine for a single exposure by splitting them across several exposures or process steps. Splitting into two is double patterning.
SADP (self-aligned double patterning)
Forming spacers on the sidewalls of a core pattern, removing the core and using the spacers as the mask to make a pitch half of what lithography alone allows. Doing it twice is SAQP.
Overlay
How accurately a new pattern is placed relative to the patterns made in earlier steps. Misalignment causes defects such as missed contacts, so it is controlled at the nanometer level.
ALD
Atomic layer deposition. A deposition method that repeats self-limiting surface reactions of precursor gases to control film thickness one atomic layer at a time.
Hard mask
A durable mask of nitride, carbon, metal or similar used when photoresist alone cannot survive the etch, for example in deep-hole etching.
Immersion lithography
Exposure with the space between the lens and the wafer filled with water. It allows a numerical aperture above 1, so ArF excimer lasers can form finer patterns.
EUV lithography
Lithography using extreme ultraviolet light with a wavelength of 13.5 nm. It uses reflective optics and a vacuum environment.

Related comparisons