NanoWiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== General ======
A resist is a material that changes its property upon exposure, most //e-beam resists// undergo a chemical change because of the electron radiation damage and can also be exposed by deep UV and X-ray radiation; they are insensitive to near UV and daylight.

E-beam resists must preferably be chemically resistant and not degas in vacuum, even while being bombarded with atoms or ions. They must not deform ('creep', 'swell', etc.) and should therefore be solids. Resists are usually spin-coated and must be solved in a casting solvent to become spinnable. The final thickness of a resist depends (among others) on viscosity and spinner rpm. The relation between thickness and rpm is described by a simple model but experimental data is also available: the so called //spin curve//. Spin time should be long enough for the thickness to level out/stabilize (usually >1 min). Uniformity of the coating is determined by the acceleration profile of the spin coater and by the density of dust particles on the substrate and solid particles in the resist (causing point-defects, filter the resist for best results). For the small 15x15 mm substrates we use, rpm must be high (3000-6000) and acceleration must be medium to fast (0-4 seconds with 1 spreading step at 500 rpm for a couple of seconds and at most 1 intermediate speed). Recipes 1 and 2 of the spinner in the cleanroom are 6000 rpm and 4000 rpm recipes that give quite good results, this does not mean that they are the global optimum for the parameters, feel free to experiment (but do not change recipes 1 and 2) and let me know about your findings so that you can write them up here. To make a spin coating process reproducible, it is important to control variables like air temperature, substrate temperature (do not put resist on a substrate that comes directly from the hot plate!), time between dispense and spinner start, presence of the spinner cover etc. It is also critical to use a clean spin bowl because vapours of previous resists can interfere very badly! So clean the spin bowl and also make sure that you always use clean pipet tips and that no air bubbles are present.

A developer is a chemical (usually a liquid) that solves exposed and unexposed parts of the resist with very different rates. The ratio between these rates is called selectivity and at a sufficient dose the selectivity becomes constant. The dose at which this happens is called the critical dose. An important property of a resist is how fast it solves in the developer when exposed below the critical dose, the slope of the linear approximation at low dose is called the contrast of the resist. A high contrast resist will go steeply from 'black' to 'white' while a low contrast resist has less steep edges and can be over or under developed (this allows for tunability) because the proximity effect causes the dose profile to change gradually instead of sharply. High contrast resists give robust results for very high resolutions, while sensitive low contrast resists are good for creating undercut profiles.

The critical dose of a resist is called the sensitivity (both in C/m<sup>2</sup>) in the 'jargon'. I don't like this word because a sensitive resist has a low sensitivity, according to this definition (but Elphy uses it). The sensitivity depends strongly on beam HV and weakly on thickness (at low HV the electrons have more interactions in the resist so you need less of them, which lowers the critical dose, makes the resist more sensitive and therefore lowers the sensitivity!). At low beam HV the forward scattering radius will increase, lowering the resolution (the performance of the electron-optical column is also worse).

Do a //dose test// to determine the sensitivity for your substrate, thickness etc. Resist thickness can be estimated from the spin curves available in the clean room.

====== Positive resists ======
Definition: A positive resist is a resist of which the exposed parts develop in a developer.
(this is of course a bad definition because this is a property of the process and not of the resist itself, but it is part of the 'jargon'. PMMA is positive and negative, it depends on the dose and the developer)

Most positive resists work by the principle of 'chain scission': they consist of a polymer with a well defined chain mass (low standard deviation of the chain mass) and the exposure process (e-beam or deep UV) cuts the chains in short fragments. The developer solves the short fragments much faster than the long chains so that the exposed resist will be developed away and the unexposed resist remains.

Many positive resiste are very selective and can be developed several times. For critical applications it is therefore easily possible to do a dose test on the resist layer that you are working with!

==== PMMA ====
Do not heavily overexpose, PMMA will become negative at 20x overexposure! 80 µC/m<sup>2</sup> is a good starting value for the sensitivity. The spin curve is plotted below. The datasheet can be found here.

The process for 950k PMMA A4 (average chain mass is 950000 <sup>12</sup>C atoms, 4 mass % solid solved in Anisole) is as follows:

  - Clean substrate with DI-water sprayer (use high glass)
  - Dry, N<sub>2</sub>
  - Clean substrate with IPA, 1 min ultra sound (use low glass)
  - Dry, N<sub>2</sub>
  - Bake, 120 C, 2 min
  - Cool substrate
  - Put hotplate at 160 C
  - Take a clean pipet tip
  - Open PMMA bottle, filter if necessary
  - Transfer 1 ml PMMA in the tip (do not suck it into the pipet! **clean pipet very well by sucking in acetone if this happens**, clean glass used for cleaning pipet.)
  - Push a little bit back in the bottle to avoid bubbles
  - Dispense, cover the substrate fully, optionally put cover on spinner, push ''start''
  - Immediately put leftover PMMA back in bottle, throw away tip and close bottle (keep bottle open time to a minimum because solvent evaporation will change viscosity and therefore spin curve)
  - Bake, 160 °C, 30 min (process is tolerant for changes in this time)
  - Clean spinner bowl and chuck in the meanwhile
  - Expose, 80 µC/cm<sup>2</sup>
  - Develop (accurately) 35 sec in PMMA developer\\
High contrast developer (for high resolution): MIBK:IPA 1:3. \\
High sensitivity developer (for the impatient): MIBK:IPA 1:1.

There are dedicated measure glasses for diluting PMMA, **never use them for something else**.

==== MMAxMAA (PMMA copolymer) ====
Copolymer resists are based on a mixture of PMMA and x % MAA (x is usually 8.5). Read the MSDS for MAA before starting. P(MMA co MAA) is a sensitive, low contrast resist that fits well in the PMMA chemistry and can be developed in one go with PMMA. It is usually used in a resist multilayer.
A dose of 80 uC/m<sup>2</sup> will result in an intermediate undercut.

The process for copolymer MMA(8.5)MAA EL y (y is the solid fraction) in ethyl lactate is as follows:

  - Clean substrate with DI-water sprayer (use high glass)
  - Dry, N<sub>2</sub>
  - Clean substrate with IPA, 1 min ultra sound (use low glass)
  - Dry, N<sub>2</sub>
  - Bake, 120 C, 2 min
  - Cool substrate
  - Put hotplate at 150 C
  - Take a clean pipet tip
  - Open copolymer bottle, filter if necessary
  - Transfer 1 ml copolymer in the tip (do not suck it into the pipet! **clean pipet very well by sucking in acetone if this happens**, clean glass used for cleaning pipet.)
  - Push a little bit back in the bottle to avoid bubbles
  - Dispense, cover the substrate fully, optionally put cover on spinner, push ''start''
  - Immediately put leftover copolymer back in bottle, throw away tip and close bottle (keep bottle open time to a minimum because solvent evaporation will change viscosity and therefore spin curve)
  - Bake, 150 C, 2 min (better: 140 C, 30 min)
  - Clean spinner bowl and chuck in the meanwhile
  - Expose, 80 uC/cm<sup>2</sup>
  - Develop (accurately) 35 sec in PMMA developer

There are dedicated measure glasses for diluting copolymer, **never use them for something else**.

==== PMGI ====
PMGI has the property that it can be easily combined chemically with PMMA type resists without mixing (photoresists typically do not allow this) but that the development is very different: it is developed in a typical alkali photoresist developer. PMGI is sensitive and has low contrast, making it very suited as a large-undercut resist. The PMMA developer does not develop the PMGI and the PMGI developer does not develop the PMMA! This allows you to inspect the sample and tune the undercut by under/over developing the PMGI.

PMGI is solved in cyclopentanone and THFA. FIXME: i'm not sure about the cyclopentanone.\\ 
Read the MSDS's before starting!

The process for PMGI SF5 is as follows:

  - Clean substrate with DI-water sprayer (use high glass)
  - Dry, N<sub>2</sub>
  - Clean substrate with IPA, 1 min ultra sound (use low glass)
  - Dry, N<sub>2</sub>
  - Bake, 120 °C, 2 min
  - Cool substrate
  - Put hotplate at 190 °C
  - Take a clean pipet tip
  - Open PMGI bottle, filter if necessary
  - Transfer 1 ml PMGI in the tip (do not suck it into the pipet! **clean pipet very well by sucking in cyclopentanone if this happens**, clean glass used for cleaning pipet.)
  - Push a little bit back in the bottle to avoid bubbles
  - Dispense, cover the substrate fully, optionally put cover on spinner, push ''start''
  - Immediately put leftover PMGI back in bottle, throw away tip and close bottle (keep bottle open time to a minimum because solvent evaporation will change viscosity and therefore spin curve)
  - Bake, 190 °C, 60 min
  - Clean spinner bowl and chuck in the meanwhile
  - Expose, 160 µC/cm<sup>2</sup>
  - Develop 3 min in PMGI 101 developer


===== Bi-layers =====
Bi-layers or multilayers are needed for producing undercut patterns and other tricks I will not mention here. These recipes should be a good starting point:

==== PMMA/PMGI ====
This is a good choice for a liftoff mask for sputtering processes because sputtering requires a large undercut profile in the resist. Spin PMGI, bake, spin PMMA on top and bake again. A good starting point for the dose is 180. Development of the 2 resist layers is fully independent: develop PMMA first (see above), PMMA developer leaves PMGI untouched. Subsequently develop PMGI (PMGI developer leaves PMMA untouched).

==== PMMA/MAA ====
PMMA/MAA is also a resist stack that produces good undercut, but somewhat smaller than PMGI/PMMA. The process is also less controllable but development is in 1 step (PMMA developer develops MAA).
Use a development time of 40 seconds.

====== Negative resists ======



==== Ma-N 24xx ====
Ma-N 24xx from MicroChem is a very easy to use negative resist with a sensitivity of about 160?C/cm<sup>2</sup>:
OLD RECIPE:
  - Spin Ma-N
  - Bake 90 °C, 12 min
  - Expose 140 C/cm<sup>2</sup> **(NOW WRONG SEE BELOW - Mar. 2006)**
  - Develop, Ma-D, 30-40 sec 
  - Descum, O<sub>2</sub>, 0.1 mbar, 100W, 2 min

The last step is important since Ma-N often leaves a bit of residue.

**__NEW RECIPE__**(Since March 2006)

With the newer developer the exposure dose is now 45, and the develop time just over 1 minute. This new developer is meant to remove issues of residues (they have added a surfacant) - and it is at least less visible than before, so the descum is perhaps unnecessary (?).

==== PMMA ====
PMMA exposed with a very high dose (>20x the normal clearing dose) crosslinks and is not removable by acetone. This gives the ultra high resolution process:

  - Coat PMMA and bake (see above)
  - Expose at a very high dose
  - Develop in acetone

Resolutions of 10nm can be achieved by this process. Crosslinked PMMA also serves as a good etch mask. See this paper about crosslinking PMMA before ion etching.

==== HSQ ====
Hydrogen silsesquioxane (available from Dow Corning) works as a negative resist. Use the recipe:

  - Spin coat HSQ:MIBK (1:1), 2000 rpm (40-50 nm)
  - Bake 150 °C 2 min, 220 °C 2 min
  - Dose and proximity test: 500-1000 µC/cm<sup>2</sup>
  - Develop 2 min MIF312:H2O (1:1), 15 sec MF322:H2O (1:9), 15 sec H2O

==== AZ 5214 ====
AZ 5214 is typically used as an optical resist that can be made negative by an image reversal step consisting of a reversal bake and a flood exposure. This works with e-beam exposure as well:

  - Spin resist (see [[optical lithography]])
  - Expose with 80 µC/cm<sup>2</sup>
  - Reversal bake
  - Flood expose (15 sec with mask aligner)
  - Develop (see [[optical lithography]])

====== Resist stripping / liftoff ======
  - PMMA and PMMA/MMA double layer get stripped well with acetone
  - PMMA/PMGI is stripped by MNP (red-cap bottle in sputtering room)
  - MaN1410 is removed extremely well with acetone
  - If your resist has become resistant to any solvent, 20 min oxygen etch in Plasmalab90 does the job nicely


----

====== Recipes summary ======
The parameters in the following recipes are used by most users. The recipes can be used as a guideline, but should be checked for your specific process. \\
Some resists can be used for e-beam and photolithography.

===== Positive Resist recipes =====

==== AR-P 617.03 (Copolymer) ====
  * spin at 4000rpm
  * bake at 200°C for 10 min

==== AR-P 642.06 ====
  * spin at 4000 rpm
  * bake at 150 °C for 3 min

==== AR-P 662.04 ====


==== AR-P 662.06 ====
  * spin at 4000rpm
  * bake at 150ºC for 3 min
  * e-beam 30kV, 180 µC/cm²
  * develop MIBK/IPA (1:3) 30 sec
  * stop IPA, 10 sec

==== AR-P 672.045 (PMMA 950K) ====
  * spin at 4000rpm
  * bake at 180ºC for 2 min
  * e-beam 30kV, 300 µC/cm² 
  * develop MIBK/IPA (1:3) 30s
  * stop IPA 30s

==== AR-P 6200.09 (CSAR 62) ====

  * Bake 150 °C, 60sec hotplate
  * e-beam exposure 65 µC/cm²
  * develop AR 600-546 60 sec
  * stopper AR 600-60 30sec , or IPA  
  * remover: AR 600-71, AR-300-76

===== Negative Resist recipes =====

==== AR-N 7500.08  ====

==== AR-N 7500.18 ====

==== AR-N 7700.08 ====
  * spin at 4000rpm
  * bake at 85ºC for 1 min
  * e-beam 20kV at least 24 µC/cm²
  * prebake 100ºC for 2 min
  * develop AR300-47 30s
  * stop demiwater 30s 

==== AR-N 7700.18 ====

===== Other =====

==== Conductive Polymer PEDOT:PSS ====
  * spin at 4000 rpm
  * let it dry a few minutes, typically you need three layers
  * rinse with water before developing
  * 
==== Electra92 ====
  * spin at 4000 rpm
  * one layer should be enough
  * bake at 105 °C for  5min on hot plate
  * rinse with water before developing

===== Optical Resists ======

^  Resist   ^  Spincoater     ^  Baking         ^  Developing  ^  Remarks  ^
|  AZ5214E  |  4000/6000rpm   |  90 C, 2.5 min  |  50% AZ312MIF + 50% H<sub>2</sub>O, 40 s  |  -  |
|  HPR205   |  4000/6000rpm   |  90 C, 2.5 min  |  50% AZ312MIF + 50% H<sub>2</sub>O, 40 s  |  -  |
|  MaN1410  |  3000rpm,30-45s |  90 C, at least 90s  |  MaD533, 15 s  |  negative photoresist  |
|  MaP1205  |  4000/6000rpm   |  100 C, 30 s  |  90% MaD532 + 10% H<sub>2</sub>O, 30 s  |  100% MaD532 can also be used, takes 10-15 s  |