Specific Process Knowledge/Etch/KOH Etch/ProcessInfo: Difference between revisions

From LabAdviser
Kabi (talk | contribs)
Kabi (talk | contribs)
No edit summary
 
(44 intermediate revisions by 5 users not shown)
Line 1: Line 1:
'''Feedback to this page''': '''[mailto:labadviser@danchip.dtu.dk?Subject=Feed%20back%20from%20page%20http://labadviser.danchip.dtu.dk/index.php/Specific_Process_Knowledge/Etch/KOH_Etch/ProcessInfo click here]'''
'''Unless anything else is stated, everything on this page, text and pictures are made by DTU Nanolab.'''
'''All links to Kemibrug (SDS) and Labmanager Including APV and QC requires login.'''
==QC==
==QC==
===Quality Control (QC) for the KOH Si etching baths.===
===Quality Control (QC) for the KOH Si etching baths.===
 
'''All links in the table below reguire login to labmanager'''
{| border="1" cellspacing="2" cellpadding="2" colspan="3"
{| border="1" cellspacing="2" cellpadding="2" colspan="3"
|bgcolor="#98FB98" |'''Quality Control (QC) for Si Etch 01, and Si Etch 02'''
|bgcolor="#98FB98" |'''Quality Control (QC) for Si Etch 01, and Si Etch 02'''
|-
|-
|
|
*[http://labmanager.danchip.dtu.dk/d4Show.php?id=3203&mach=9 The QC procedure for Si Etch: 01]<br>
*[http://labmanager.dtu.dk/d4Show.php?id=1565&mach=248 The QC procedure for Si Etch: 01 and Si Etch 02: KOH]<br>
*[http://labmanager.danchip.dtu.dk/d4Show.php?id=1565&mach=248 The QC procedure for Si Etch: 01]<br>
*[http://labmanager.dtu.dk/d4Show.php?id=5476&mach=407 The QC procedure for Si Etch: 03]<br>
*[http://labmanager.danchip.dtu.dk/view_binary.php?type=data&mach=49 The newest QC data for KOH2]<br>
*[http://labmanager.danchip.dtu.dk/view_binary.php?type=data&mach=248 The newest QC data for KOH3]<br>
{| {{table}}
{| {{table}}
| align="center" |  
| align="center" |  
Line 25: Line 30:
|Time
|Time
|90 min
|90 min
|-
|Elevator
|On
|-
|-
|Substrate
|Substrate
Line 30: Line 38:
|-|-
|-|-
|Masking
|Masking
|No masking
|Thermal SiO<sub>2</sub>
|-
|-
|}
|}
Line 36: Line 44:
{| border="3" cellspacing="1" cellpadding="2" align="center" style="width:500px"
{| border="3" cellspacing="1" cellpadding="2" align="center" style="width:500px"
!QC limits
!QC limits
!Si Etch 01
!Si Etch 1
!Si Etch 02
!Si Etch 2
!Si Etch 3
|-
|-
|Etch rate in Si(100)
|Etch rate in Si(100)
|1.3 ± 0.1 µm/min
|Latest QC Data on [http://labmanager.dtu.dk/function.php?module=Machine&view=view&mach=376 Si Etch 1 Info] page (under Documents)
|1.29 ± 0.06 µm/min
|Latest QC Data on [http://labmanager.dtu.dk/function.php?module=Machine&view=view&mach=390 Si Etch 2 Info] page (under Documents)
|Latest QC Data on [http://labmanager.dtu.dk/function.php?module=Machine&view=view&mach=407 Si Etch 3 Info] page (under Documents)
|-
|-
|Roughness
|Roughness
| not measured
| Only evaluated visually
| not measured
| Only evaluated visually
| Only evaluated visually
|-
|-
|Nonuniformity
|Nonuniformity
|< 3%
|< 3%
|< 3%
|< 3%
|< 3%
Line 60: Line 72:
==Mixing KOH==
==Mixing KOH==


Danchips standard solution in the KOH baths has a concentration of 28 wt% and is mixed of either 500 g KOH pills to 1 l of water or 1liter of a premade 50wt% KOH solution to 1.2 l of water.
Nanolabs standard solution in the KOH baths has a concentration of 28 wt% and is mixed of 1 liter of a premade 50wt% KOH solution to 1.2 l of water.


If you want to make your own mixture of KOH in a beaker in Fumehood 6 you can either do this by using KOH pills or a premade mixture of 50wt%. Using the pills the mixture will spontaneously heat up when mixing which is not the case when using the premade mixture.


If you want to make solution using the KOH pills you must remember that the pills are only 85% pure.
If you use the 50wt% premade KOH solution you can calculate the amount of solution X you need to add 1 liter of water to achieve a concentration of Y wt% by:


For 1 l of water you can calculate the required amount of KOH pills (X) to achieve the desired concentration in wt% (Y) by:


<math>\frac{(0.5\cdot 1.509g/ml\cdot X)}{(X\cdot 1.509g/ml+1000g)}=Y</math>


(0.85wt%∙X)/(X+1000g)=Y


<math>X=\frac{(Y\cdot 1000g)}{(1.509g/ml\cdot(0.5-Y))}</math>


X=(Y∙1000g)/(0.85wt%-Y)
So for a 10wt% solution you will need 133 g pills for each liter of water.
If you use the 50wt% premade solution you can calculate the amount of solution X you need to add 1 liter of water to achieve a concentration of Y wt% by:
(0.5wt%∙1.509g/ml∙X)/(X∙1.509g/ml+1000g)=Y


where 1.509 g/ml is the density of a 50wt% solution KOH at 20 °C.


X=(Y∙1000g)/(1.509g/ml∙(0.5-Y))
So for a 10 wt% solution you will need 165 ml of KOH (50wt%) to 1 l of water.


== Backside protection ==


where 1.509 g/ml is the density of a 50wt% solution KOH at 20 °C.  
'''It is possible to protect the backside of the wafer against the KOH etch using special wafer holder.'''


So for a 10 wt% solution you will need 165 ml of KOH to 1 l of water.
Nanolab has backside protection holders for the two baths KOH 2 and KOH 3 for 4” wafer, and it is possible to mount two wafers in one holder. The material is PEEK, which is suitable for nearly all etchants and you will find the holders by the KOH baths. The recommended space from the wafer edge to your pattern is 7 mm. The thickness of the wafer would typically be 350 μm and the wafer would typically be polished on both sides, however with the right gasket (contact wet chemistry) the holders can accommodate 500-525 μm wafers. In order to avoid differences in pressure between the backside of the wafer and the surrounding environment, the venting tubes should be filled with water. The holder is clamped around the wafers by screws, if these screws are tightened too much or too little, it will affect the etching result. For the KOH 2 bath we have another type of holder which is clamped around the wafer with clamps (it cannot be adjusted to another wafer thickness).
See more specific details about the screwed holders [https://ammt.com/support/files/AMMT_PI_WETandem.pdf here], the holders we have at Nanolab is the Tandem4 type.


==Theory==
==Theory==
Line 96: Line 101:
Due to the almost inert (111)-planes it is possible by KOH etching to realize high aspect ratio structures in sigle crytalline silicon using the (111)-planes as sidewalls. In Si(100) these sidewalls are inclined - 54.7<sup>o</sup> with respect to the (100) surface - whereas in Si(110) the sidewalls are vertical (see figures below).
Due to the almost inert (111)-planes it is possible by KOH etching to realize high aspect ratio structures in sigle crytalline silicon using the (111)-planes as sidewalls. In Si(100) these sidewalls are inclined - 54.7<sup>o</sup> with respect to the (100) surface - whereas in Si(110) the sidewalls are vertical (see figures below).


<gallery caption="Anisotropic wet silicon etch: dependency on crystal orientation" widths="380px" heights="150px" perrow="2">  
<gallery caption="Anisotropic wet silicon etch: dependency on crystal orientation" widths="760px" heights="150px" perrow="1">  
Image:KOH_Anisotropy.jpg|Etched profile when etching Si(100).
 
Image:KOH_Anisotropy(110).jpg|Etched profile when etching Si(110).
Image:KOH-etch.JPG|Etched profile when etching Si(100) and when etching Si(110) respectively
 
<!-- Image:KOH_Anisotropy.jpg|Etched profile when etching Si(100).-->
<!-- Image:KOH_Anisotropy(110).jpg|Etched profile when etching Si(110).-->
</gallery>
</gallery>


<br clear="all"/>
<br clear="all"/>
For Si(100), the relation between the width of the bottom of the etched groove (W<sub>b</sub>) and the width of the opening (W<sub>o</sub>) at the wafer surface in a groove etched to the depth l is given by:
For Si(100), the relation between the width of the bottom of the etched groove (W<sub>b</sub>) and the width of the opening (W<sub>o</sub>) at the wafer surface in a groove etched to the depth d is given by:




<math>W_b = W_o - 2lcot(54.7^o) = W_o - \sqrt{2} l</math>
<math>W_b = W_o - 2d cot(54.7^o) = W_o - \sqrt{2} d</math>




Line 112: Line 120:
===Definition of <110> alignment structures===
===Definition of <110> alignment structures===


The etch rate dependence on the crystallographic planes can be used to determine the <110> crystal directions with high precision (better than +/- 0.05 <sup>o</sup>). A fast method for doing this, using the symmetric under-etching behavior around but not at the <110>-directions, was described by Vangbo and Bäcklund in J. Micromech. Microeng.'''6''' (1996), 279-284. High-precision control of the <110>-direction during alignment can be necessary in order to control the dimensions of KOH-etched structures (e.g. precise control of V-groove dimensions). A dedicated mask (MASK NAME) has been designed for this purpose.
The etch rate dependence on the crystallographic planes can be used to determine the <110> crystal directions with high precision (better than +/- 0.05 <sup>o</sup>). A fast method for doing this, using the symmetric under-cutting behavior around but not at the <110>-directions, was described by [https://iopscience.iop.org/article/10.1088/0960-1317/6/2/011/meta Mattias Vangbo and Ylva Bäcklund 1996 J. Micromech. Microeng. '''6''', 279-284]. High-precision control of the <110>-direction during alignment can be necessary in order to control the dimensions of KOH-etched structures (e.g. precise control of V-groove dimensions).  


<br clear="all" />
<br clear="all" />


===Etch rates: Empirical formula (Seidl et al)===
<!-- ===Etch rates: Empirical formula (Seidl et al)===


The following empirical formula can be used for concentrations in the range of 10-60 wt%:
The following empirical formula can be used for concentrations in the range of 10-60 wt%:
Line 124: Line 132:
where k<sub>0</sub> = 2480 µm/hr (mol/l)<sup>-4.25</sup>, E<sub>a</sub> = 0.595 eV for Si(100)
where k<sub>0</sub> = 2480 µm/hr (mol/l)<sup>-4.25</sup>, E<sub>a</sub> = 0.595 eV for Si(100)


and  k<sub>0</sub> = 4500 µm/hr (mol/l)<sup>-4.25</sup>, E<sub>a</sub> = 0.60 eV for Si(110)
and  k<sub>0</sub> = 4500 µm/hr (mol/l)<sup>-4.25</sup>, E<sub>a</sub> = 0.60 eV for Si(110) -->

Latest revision as of 14:06, 27 March 2023

Feedback to this page: click here

Unless anything else is stated, everything on this page, text and pictures are made by DTU Nanolab.

All links to Kemibrug (SDS) and Labmanager Including APV and QC requires login.


QC

Quality Control (QC) for the KOH Si etching baths.

All links in the table below reguire login to labmanager

Quality Control (QC) for Si Etch 01, and Si Etch 02
QC Recipe:  
Solution 28 wt% KOH
Temperature 80°C
Time 90 min
Elevator On
Substrate Si (100)
Masking Thermal SiO2
QC limits Si Etch 1 Si Etch 2 Si Etch 3
Etch rate in Si(100) Latest QC Data on Si Etch 1 Info page (under Documents) Latest QC Data on Si Etch 2 Info page (under Documents) Latest QC Data on Si Etch 3 Info page (under Documents)
Roughness Only evaluated visually Only evaluated visually Only evaluated visually
Nonuniformity < 3% < 3% < 3%


Mixing KOH

Nanolabs standard solution in the KOH baths has a concentration of 28 wt% and is mixed of 1 liter of a premade 50wt% KOH solution to 1.2 l of water.


If you use the 50wt% premade KOH solution you can calculate the amount of solution X you need to add 1 liter of water to achieve a concentration of Y wt% by:




where 1.509 g/ml is the density of a 50wt% solution KOH at 20 °C.

So for a 10 wt% solution you will need 165 ml of KOH (50wt%) to 1 l of water.

Backside protection

It is possible to protect the backside of the wafer against the KOH etch using special wafer holder.

Nanolab has backside protection holders for the two baths KOH 2 and KOH 3 for 4” wafer, and it is possible to mount two wafers in one holder. The material is PEEK, which is suitable for nearly all etchants and you will find the holders by the KOH baths. The recommended space from the wafer edge to your pattern is 7 mm. The thickness of the wafer would typically be 350 μm and the wafer would typically be polished on both sides, however with the right gasket (contact wet chemistry) the holders can accommodate 500-525 μm wafers. In order to avoid differences in pressure between the backside of the wafer and the surrounding environment, the venting tubes should be filled with water. The holder is clamped around the wafers by screws, if these screws are tightened too much or too little, it will affect the etching result. For the KOH 2 bath we have another type of holder which is clamped around the wafer with clamps (it cannot be adjusted to another wafer thickness). See more specific details about the screwed holders here, the holders we have at Nanolab is the Tandem4 type.

Theory

Definition of structures

Due to the almost inert (111)-planes it is possible by KOH etching to realize high aspect ratio structures in sigle crytalline silicon using the (111)-planes as sidewalls. In Si(100) these sidewalls are inclined - 54.7o with respect to the (100) surface - whereas in Si(110) the sidewalls are vertical (see figures below).


For Si(100), the relation between the width of the bottom of the etched groove (Wb) and the width of the opening (Wo) at the wafer surface in a groove etched to the depth d is given by:




Definition of <110> alignment structures

The etch rate dependence on the crystallographic planes can be used to determine the <110> crystal directions with high precision (better than +/- 0.05 o). A fast method for doing this, using the symmetric under-cutting behavior around but not at the <110>-directions, was described by Mattias Vangbo and Ylva Bäcklund 1996 J. Micromech. Microeng. 6, 279-284. High-precision control of the <110>-direction during alignment can be necessary in order to control the dimensions of KOH-etched structures (e.g. precise control of V-groove dimensions).