THE CHEMICAL ENGINEERS' RESOURCE PAGE
EMAIL ARCHIVE MESSAGE 032   

I am trying to find articles on "scale-up", i.e. going from benchtop to plant.  Do you know of any articles that might describe common problems or best practices?

     There are three articles you should try to get a hold of that may help.

1. "Pharmaceutical Process Development is Different", CEP, Sept 1998
2. "Achieve the Right Balance in Pharmaceutical Pilot Plants", CEP Feb 1998
3. "Pilot Plants Confirm Process Validity", Chemical Engineering Magazine, June 1998

The three articles talk about scale up and should help. Even if you are not in the pharmaceutical business, the first two articles can offer a great deal of insight on scale up even for chemical/perochemical processes.

In general, scaling up can be rather straight forward. However, many things just don't scale directly. I am reminded of a story told to me by one of my former bosses. In his younger days, he was working on a bench scale process that involved a unique reactor design (can't remember details). The reactor was tubular and the engineers got all the data necessary; velocities, residence time, temperatures, pressures, etc. They scaled the thing up to pilot plant size and it didn't work. They tried everyting they could think of. Even though it should have worked, it didn't. So, what they ended up doing was taking multiple "small" scale reactors and installing them in a
large shell; sort of like a shell-and-tube heat exchanger. That did the trick!

Some of the biggest problems facing scale-up is velocity profiles of liquids and gases. Even if you get the right mass flow rate, the larger size tubes, vessel, pipe, whatever, may not give you just the right distribution pattern you need or that you found on the bench scale. Mixing is another problem. It is easy to mix on a bench. It is another matter when you need to use an agitator in a large vessel. You may get many "dead" spots that are not apparant on the bench. The degree of mixing may be so much that it becomes a cost nightmare in terms of both capital and operating in a full size plant versus the bench. Separation is another problem. It's pretty easy to get
good distribution between liquid/vapor (distillation, stripping, absorption etc.) or between liquid/liquid as in extraction on the bench but it may be another story in a 3, 4, 6 etc. foot column. Is this pressure or vacuum? It's pretty easy to maintain a good vacuum on a bench but doing it on a large scale is another story. Pipe flanges and flanges between pipe and equipment or equipment and equipment (agitator on top of a vessel) can and often does leak making it difficult to maintain vacuums.

What about that jerry rigged equipment you came up with to make the process work on the bench. Now try to get an industrial size of the same gizmo!

VLE data can be a nightmare. I worked on a project some years back where the client performed some experiements in stripping benzene from waste water. The vessel was a 55 gal drum and steam was sparged in. They never measured the VLE or stripping efficiency. They then expected us to design a tank (can't remember but I think it was in the order of 500,000 gal) with a steam
sparger to strip benzene off before the liquid was sent for further downstream processing. NO WAY JOSE!  These are some of the things you will need to worry about.

People doing bench scale work forget that Process Engineers need VLE data to size columns or other separation equipment and so don't get the data or don't get enough data. What about safety? Is your reaction capable of running away? It's easy to control that on the bench but not in a 5,000 gal reactor vessel. Do you have enough VLE/enthalpy data to size relief devices (relief valves and/or rupture disks)?

Can you pump the stuff? It's easy to transfer reagents, intermediates and products from one beaker to another (you can pour it, even if it is done under a hood). In the big world, can you pump it?

Phil Leckner
First Content Manager
www.cheresources.com