The 8 Shot

Parr Instrument Company 4749 General Purpose Acid Digestion Vessel

This is a Parr Instrument Company acid digestion vessel; however, people typically refer to them as “Bombs”. This is due to their modus of operandi having to remain in oven at high temperatures (~200C), with aqueous or organic solvents contained inside. If the temperature is left unchecked and the pressure exceeds the rating, the Teflon liner inside will rupture or explode.

In my research lab at San Francisco State University, led by Dr. Andrew S. Ichimura, we investigate titanium dioxide (TiO₂) as a photocatalyst. Specifically, we produce thin films, varying the synthesis parameters, and characterize their unique properties. This involves the morphology, size and catalytic performance of the anatase TiO₂ grains produced. Inevitably, the goal of our research is to find the properties of the thin films and the subsequent synthesis parameters that optimize their photocatalytic performance.

Anatase TiO₂ Single Crystal with (001) and (101) facets labeled (left) and Atomic Force Microscopy (AFM) scan of (001) textured anatase thin film (10 μm x 10 μm).

The acid digestion vessel we use has been the industry standard for quite some time for nanocrystal synthesis. Often used in the production of zeolites, perovskites, metal oxide nanoparticles and even quantum dots. It works great for said production of nanoparticles, however, it can be quite inefficient for thin film research.

Due to the large volume, in comparison to the size of our substrates (1”x 5/16”), there is a significant waste in terms of precursors used. Furthermore, an undesirable amount of nucleation occurs forming large clumps of crystals, referred to as spherulites, that land on the surface of the film.

However, the largest limitation for our thin film research is being constrained to performing one synthesis per vessel. In the process of producing a series of samples, such that we can compare their properties, we must thoroughly clean each Teflon liner between syntheses. This involves a day of baking the liner with hydrofluoric acid (HF), to dissolve any left behind particles, followed by a water bake to remove the adsorbed HF. This process, including syntheses, takes at least 2 days to complete.

The only way to increase throughput is using multiple reaction vessels, which cost thousands of dollars each, inducing uncertainty that each vessel experienced the same time and temperature during synthesis. To resolve this inaccuracy in our science, increase throughput, and reduce waste, I started the process of engineering a multi-vessel reactor.

X-Ray Fluorescence Spectroscopy (XRF) of acid digestion vessel components (top and chamber).

To start this process, I reverse engineered the 4749 acid digestion vessel we typically use. To understand what is used to prevent corrosion of the vessel, and galling of the threads, I performed XRF on all the components in the assembly. This revealed that the top is a proprietary nickel alloy, to provide corrosion resistance and eliminate thread galling, and all other components are 304 stainless steel. I now understood how they were able to utilize a threaded design, however, I am left unable to replicate it due to the high cost of such a nickel alloy ($$$$).

Example of items that utilize a bayonet style engagement (Electrical cables, Automative light bulbs, and a bayonet).

I initially was clouded by the idea of trying a threaded design. However, I was heavily encouraged not to pursue by SFSU machinist, and my mentor, Peter Verdone. Upon researching different types of engaging parts, I came across the bayonet style connector. This has multiple advantages. It offers quick and simple engagement, as opposed to threading the cap on for a whole minute. More importantly, it enables loose clearances for ease of manufacturing and eliminates any galling during use.

Cross section of Teflon liner used in Parr Instrument Company 4749 Acid Digestion Vessel

While I was left to contemplate the vessel design, I began working on how to produce a Teflon liner. Due to the nature of Polytetrafluoroethylene (PTFE), it is very slippery and does not like to seal. This is circumvented by applying a large amount of pressure to the mating surfaces during use. Although, thermal expansion must still be accounted for.

This instance the design exemplifies the need to design around the manufacturing tools at my disposal. Since I will only have access to a manual mill and lathe, the geometry of this part makes manufacturing quite a tedious task. This led me to rethink what is important in terms of getting this liner to seal and allowing for ease of production.

An easier design to produce would be a straight, hollow cylinder. Neglecting the lipped design of the original product. This would involve a cap shaped like top hat which retains the horizontal and flat mating surfaces, while introducing a stopper to plug the inside of the cylinder. To materialize my ideas, I began learning solid modeling in SOLIDWORKS.

Drawing of single chamber, bayonet style reaction vessel with small liner design.

I completed the final design of the single chamber prototype after optimizing various attributes in the model. The overall design is largely influenced by the original “Bomb”. The only changes include the bayonet engagement top, and Teflon liner redesign in which it decreases from a 23 ml volume to a 2 ml volume. I produced prints for each part and manufactured all components by hand.

Manufactured components of single vessel bayonet prototype.

After making the prototype, I needed to test its functionality and fix any issues with its design. To do this I need to complete water loss measurements to ensure that the Teflon liner is sealing after a cycle in the oven. I begin by filling the liner with about 2 ml of water and measuring its weight. I then put it in the oven overnight at 180C. The next day I take it out, let it cool down, and measure its weight again.

The first water loss measurement I completed was a total failure. All the water in the liner was gone. However, a unique part of the design is the integration of the pressure plate. It allows for even distribution of pressure from the spring on the lid of the liner. Furthermore, I can easily remake the disk to any thickness I desire and effectively increase the spring pressure applied to the liner.

So, I began increasing the thickness in small increments until water loss became negligible. I updated my model in SOLIDWORKS with the thickness of the working pressure plate to figure out how much I am compressing the spring. This turned out to be around 300 thousandths of an inch. I then use the known spring constant of the spring, 57lbf/in, and calculate how much force is sealing the Teflon liner, 16.9lbs.

AFM scans of films from Parr reactor (left) and Custom reactor (right) using same solution.

Optical microscope images of films from Parr reactor (left) and Custom reactor (right) with white blobs representing spherulites on the surface.

After successfully getting the liner to seal, I proceeded to perform actual comparisons using thin film synthesis. The films produced appear very similar in an AFM scan, however, when observed in an optical microscope the difference is clear. The custom vessel greatly minimizes the amount of excess nucleation, due to having less moles of precursor in the same vessel, despite the exact same concentration.

Having confirmed that the bayonet and Teflon liner design works as intended, I can progress towards a higher throughput design. I decided upon a reactor that house eight reaction vessels, as they would fit in a radial array of 4” stainless steel round stock, a constraint imposed by our small ovens in the lab.

With the knowledge that the liner only requires around 17lbs of pressure to seal, I could find a more tuned spring that better suits this application. I played around with the idea of using one big spring to seal all the liners, however, I was soon reminded of the stool example. When one takes a stool, or anything with four legs for that matter, they will notice that the object does not lay perfectly flat. One leg will not be in contact with the ground because it only takes three limbs to not fall over, and most objects are not made perfectly. I realized this same logic applied to the single spring design. At least three vessels will be suspending the spring and taking all the pressure, due to imperfect tolerances in machining, and at least five vessels will not have any sealing pressure, until thermal expansion takes over.

The simplest solution to make sure all the vessels are experiencing pressure is to give them each their own spring. For this task I sought to use a special spring called and interlaced wave spring with shim ends. This type of spring allows for a low-profile design while distributing even pressure across the whole surface of the Teflon liner. Such that the height of the 8 Shot reactor is lower than either the Parr reactor or the prototype.

Drawing of the 8 Shot Reactor, as produced.