AT400 Base ALD System - USP
 

The base pressure of the AT400 (inert gas purge shutoff) is < 5 mtorr using our standard pump Edwards E2M18 (12cfm at 60 Hz)

• The system is leaked checked with a ultra sensitive He leak detector (maximum sensitivity of 4 x 10-12 Torr·L/sec) at all of the metal seals
• Please also emphasize that the system has all metal seals upstream of the sample unlike the Ultratech Savanah, which has an elastomer seal that goes completely around the chamber.  The AT400 chamber is an all aluminum chamber with a SS to Aluminum explosion welded precursor input fitting, this prevents all contaminants from entering into the chamber near the substrate.  The samples are introduced on the opposite side of the precursor inputs and any diffusion of trace H²O and O² will go to the pump port before the possibility of reaching the sample.

The system will yield quicker cycle times for their low temperature process because of two main reasons:

• Like you all mentioned, we can control the amount of precursor that goes into the system much more easily using defined volume dosing vs valve time dosing especially for high vapor pressure precursors like TMA and H²O
• Second, because the system is designed around 4" and small substrates, we do not have to use as much precursor to achieve a saturating dose through the chamber (chamber surface area is low), while at the same time the chamber flow design is optimized for more efficient purging unlike the Savannah

The SW allows for almost complete control over all of the process parameters including:

• Precursor temperature (combined with the dose volume of 12mL) controls the # of moles of precursor per pulse
• # of pulses of precursor prior to full chamber purge and introduction of second precursor
• Valve open/close times for all 10 pneumatic valves in system
• High exposure mode or dynamic mode
• Exposure time
• Precursor manifold temperatures
• Chamber temperature
• Process pressure (needle valve)
• Multilayer sequences
• Save recipes
• Purge times, tolerances, safety interlock controls etc...
   

Recipe generation

• Recipes are programmed (takes < 1min) and saved directly on the large touchscreen HMI of the system
• Can store 150 recipes, but that can be expanded to >1000 if requested
• All relevant control parameters are available
• Separate PC control is not required and would actually take more time than just creating on screen using the pull down menus and direct inputs.
• For multilayer or multicomponent processing, the user selects a saved recipe for each layer and then inputs the number of sub cycles and the total number of multilayer cycles (Ex.  You could create a multiyear process with 1 nm Al²O³ followed by 1nm HfO^2, repeated 10X for a total multilayer film of 200 nm)
• PLC based system control has many advantages over lab view/windows control platforms (see next bullet)

PLC control with touchscreen interface

• Unlike windows PCs, there are no virus scans, windows updates, USB errors (common to Savannah systems) etc..
• PLCs are the standard way of running complex automation equipment in factories because they are robust and do not have complex operating systems that get in the way of actual tool operation
• Systems are never obsolete because legacy support of PLC controllers is provided.  This is not the case for windows based system control SW. 
  If ultratech decides to discontinue that tool, then in a few years the SW could cease to function correctly as windows continues to see new versions. 
  PLC SW doesn't corrupt and the HW is hot swappable and fully supported in perpetuity by the manufacturer (Unitronics)

I can provide them with some SW screen shots if they desire.

The precursor cabinet door is sealed and can be vented from the back/top by connecting to a standard 1 1/2" ID exhaust tube and connected to their house exhaust

The prototype tool is installed in the Harvard cleanroom (Center for Nanoscale Systems).  We have >30 registered users on that particular tool and >1500 unique users at the cleanroom each year.  It can certainly handle a shared user environment.
 

System maintenance

• Pump oil should be changed as needed (noticeable color change) or at least once/year
• TC pressure gauge should be changed when/if baseline pressure drifts (sometimes 6 months, sometimes 2 years)
• The chamber can be cleaned with clean room cloths with IPA when particles are observed, this type of chamber clean can continue to be effective for 1-2 years depending on the amount of film grown and whether you are mixing metals with oxides or not.  For this application it would be more on the 2-3 year time scale (oxides, very thin films)
• When the chamber needs to be fully cleaned, it can be easily removed from system and sent for CO² blast cleaning in US or in France.

Sample size

• The holder can accommodate a 4" wafer up to 5 mm thick.
• A 4" square sample will also fit in chamber, but will overhang the round holder at the corners.
• If the holder is removed and a sample is manually placed into the chamber, it can accommodate a sample 11 mm tall.
• Custom sample holders can be made as well
• Small samples in the standard holder sit in a slight depression in the middle and do not move appreciable during pump down.

Anric Technologies the company

• Though the company is small, we contract with many different companies/groups like our manufacturing partners
• Process development and films analysis takes place mostly at my lab at Harvard and through several of the installed systems at customer sites
• We have seen a very good reception to the tool in the past 18 months and we are projecting 12-15 tools at customer sites by Q3 2017.  As the company continues to see success, we will expand as appropriate.  We will not fall into the trap that Cambridge Nanotech fell into which was to expand at a pace faster than the revenue was expanding.
• Personally I started in ALD in 2001 and have been working both in academia and industry in this field ever since my education with Professor Gordon at Harvard.

System delivery

• The standard lead time is 3 months from PO, however, we typically ship at around 2 months.
   

Advantages I didn't see discussed or questioned:

• Metalorganics and counter reactants (such as H²O) are on separate manifolds so there is no potential growth of film within the precursor manifolds, just where they "see" each other in the chamber.  Savanah uses a shared manifold, so growth occurs inside and valves can fail prematurely.  We have seen this at Harvard ( we have an AT300 (3" system) and a Savannah)
• The system can be attached to a glove box with a modification to an existing glove box or a new GB and a part that can be added to the AT400 at a later time
• We can supply all precursors filled and ready to go directly to the customer often at a price that is lower than what they can find locally.  My supplier will ship directly to their lab and handle all the export/importation issues.  Even for trimethylaluminum and diethyl zinc
• Because the system is metal sealed in the hot zone, we can heat the system to 325-350C without degrading the vacuum seals since the door o-ring seal is not in the same hot zone.
• Because of the optimization of both chamber size and dosing, we utilize precursors and facilities like power and nitrogen, more efficiently than the competitors thus having the lowest COO in the business.
  • Each precursor has a separate dose volume that is filled before be emptied into the chamber except for the oxidants which share the same dose volume.  I have included a simple diagram of the plumbing and a typical pulsing sequence.
  • Samples are changed by venting system, pulling the door open by the handle on right, retrieving old sample and then placing a new sample on the holder.  Small samples can sit in the recessed portion of holder in the center and should stay put.
  • Room temperature depositions often do take a long time, but this can be mitigated by using a different oxidant, like hydrogen peroxide or ozone, which pump faster and of course by using less H²O, which our design can more easily do than the Ultratech system.  The Ultratech Savannah cannot pulse reproducibly faster than about 10-15ms, which still can provide more water vapor to the chamber than is required for saturation.
  • Though we may create new equipment that has plasma ALD capability, the AT400 will not be able to be upgraded to that king of tool.  For reference, non of the bench top or small ALD systems like a Savannah can be upgraded to full PEALD.  You have to purchase a new tool for all of the companies.