I've been reading up on aspirator pumps in preparation for building a recirculating one based on a Procon pump scavenged from a beverage carbonator. It's rated 250 psi and 100 gph (1.7 MPa and 0.10 L/s), so it's got plenty of flow rate. I'll
principally be using it as a roughing pump for vacuum.
While reading up on this, it got me thinking about a possible configuration for a recirculating aspirator. In some product literature for a
recirculating unit, I found instructions that when aspirating organic solvents, that a continuous flow of water should be used (with the overflow
drain running) and that the tank water should be discarded at the end of the run. This is presumably because of dissolved solvents. Why not take
advantage of this rather than deal with it as a problem?
My thought was to aspirate with an organic solvent, recirculating, and to use that solvent to capture the fumes coming off an evaporator (say), and
then distill to recover the aspirating solvent and the process solvent. There are two benefits from this: solvent recovery and emissions reduction.
The reason to switch from water is to increase the solubility, so the aspirating fluid acts as a trap in addition to a pump. At the exhaust end of the
pump, you'd have a cold trap to capture what vapor remained.
Ultimate pressure from an aspirator depends upon the vapor pressure of the aspirating fluid, which is why you get better performance from cold water.
A recirculating aspirator could conceivably use any available solvent, but if you restrict yourself to room temperature to avoid the cost of a
chiller, you'd want a solvent with a reasonably low vapor pressure. For low operating cost, you'd also want an OTC solvent. I looked at ethanol,
acetone, and MEK (butanone); all of these have fairly high vapor pressures at 30° C. It's about 100 mm Hg for ethanol and MEK, 300 mm Hg for
acetone. By contrast it's 32 mm Hg for water. And it's down about 9, 12, and 11 mm Hg for o-, p-, and m-xylene. So xylene seems to make a decent
aspirating fluid, as good as ice water.
As for material compatibility, my pump is brass, but they're also available in stainless. An exploded diagram of the pump seems to indicate that the
seals are gaskets and o-rings, all available (if not from Procon) in Viton.
I don't have a particular need to do this, but plenty of the participants here do. I hope someone might find this a useful idea.Sauron - 12-11-2008 at 10:45
Recirculating (usually dual) aspirator pumps are commercially available so are easy to copy. I have an Eyela and it works very well. You have the
theory down well. The colder the water, the better the vacuum. If your local tap waater is nice and cold, no problem. If not, add ice. You will need
to deal with overflow.
Your solvent recovery condensers need to be efficient (Buchi rotavap condensers are a good model, or else Dewar condensers for dry ice/acetone if your
solvents are low boiling). You do not want any solvent reaching your aspirator system or the vacuum will go to hell.
Bottom draining, vacuum isolable receiving flasks are a big help.497 - 12-11-2008 at 17:29
So, if I understand correctly, the main factors affecting an aspirator working fluid performance are vapor pressure and viscosity right?
So how would ethylene glycol work? Its vapor pressure at room temperature is far lower than cold water, about 0.1 mm Hg. It's viscosity is quite a bit
higher though, but how much of a factor is that? The other one that came to mind is DMSO, it could theoretically get down to less 1mm Hg at room temp
and has a viscosity only slightly higher than water. It would be interesting to try if you had a pump that could handle it...watson.fawkes - 12-11-2008 at 19:46
I have found, while reading up on this subject, that aspirator pumps are also called ejector pumps or jet ejector pumps. There's a different and more
recent usage of "ejector pump" to refer to a basement sewage pump. From what I've seen, "aspirator" seems to be used for water, "ejector" for other
fluids.
The suggestion of ethelyne glycol is interesting. The main difference with a higher viscosity pumping fluid is that there's more energy lost through
internal friction (that's what viscosity it) and thus the pumping fluid heats up faster. I do imagine that a heat rejection system might be necessary.
For ethylene glycol, you could use an ordinary automotive radiator on the intake to the pump. For other fluids, you'd use a cooling coil and a
heat-transfer fluid rather than direct pumping.
The dynamic viscosity of ethelyne glycol is 17.33 cP. It's density is 1.1132 g/cm<sup>3</sup>. Thus it's kinematic viscosity is 15.57 cSt.
ISO oils are graded in cSt (centistokes), so it's got the viscosity of an ISO 15 oil. That's in the light oil category, and the kind of Procon pump
I've got would handle it just fine (were I building for this purpose).
The downside of ethylene glycol is just how hygroscopic it is. I'd worry about absorption of atmospheric water vapor. I suppose the solution to that
is a valve on the intake and exhaust to isolate the system when not in use. That's a good idea anyway.497 - 12-11-2008 at 20:08
Actually lubricating oil looks like it might be a better choice. Less toxic, less reactive and not hygroscopic. Since its vapor pressure is so low you
could run it at a higher temperature, making cooling much easier. Supposedly ISO 32 oil at 40*C has a vapor pressure of 0.00004 mm Hg. Hell, you could
run it at 150*C and still get a decent vacuum! And the viscosity is only 4 times that of glycol...
As far as solvent recovery goes, if you ran your working oil a bit above the solvents boiling point, I wouldn't think there would be much trouble with
contamination, and you might be able to conveniently condense the vapors that come off. Maybe there would be a problem with the vapors dissolving into
the oil and killing the vacuum, I don't know...
Edit: Response to the post below.
Yes, I agree, but with something like a mineral oil the temperature has to get very high before the vapor pressure becomes significant. Like
over 120*C or more. I think all the heat that a pump could ever put into the system would be easily radiated away by a working fluid that hot. If for
some reason it couldn't it would be easy to use a water cooling system, but I don't think that would ever be needed. That's the beauty of using
mineral oil...
Still, I suspect there must be some problem with using it, mainly because I've never heard of it being used that way. That makes me think there must
be some other factor that keeps it from being practical.
[Edited on 12-11-2008 by 497]chemrox - 12-11-2008 at 22:27
)ne of the virtues of water in such systems is its high heat capacity . Even so, the pump motors in aspirator pumps give off heat which the water
picks up. On longer runs it is wise to either change the water continuously or add ice. With other solvents these options become difficult.octave - 13-11-2008 at 08:46
Couldn't you use dry ice as a coolant if you're using oil?497 - 13-11-2008 at 12:03
Yes you could.. But I'm not sure why you'd want to, I think it would start to get extremely viscous/gelled at than temperature, unless it was very
light oil.
Formal similarity of ejector pump and diffusion pump
watson.fawkes - 14-11-2008 at 06:59
This discussion has been useful for me. I've had a few trains of thought on it, which I'll post separately. Here's the first.
One useful insight has come to me on this subject: the formal similarity of an ejector pump and a diffusion pump.
Both use direct transfer of momentum by collision with a pumping fluid. With an ejector pump, the fluid is in the liquid
phase; with a diffusion pump, the fluid is in the vapor phase.
Both have a pumping loop. With an ejector pump, it's a rotary vane or other pump; with a diffusion pump, it's a pressure gradient form by an
evaporator and a condenser. With a water aspirator, one member of the "loop" is the infinite reservoir of water, with source and sink partly
disconnected. This changes the operation, but not very much the analysis.
Both have a heat pump. With a diffusion pump, this is the how the pumping loop operates. With an ejector pump, this is used to dissipate the
heat generated by mechanical pumping and the heat picked up from the environment.
I also have an idea about why mineral oil may not be common in these pumps. Up until pretty recently, water was not considered scarce; there
wasn't much need to economize on water usage. When the switch happened, there were already existing solutions, and so folks stayed with what was known
to work.
As to why an oil-based ejector pump might not be appropriate, there's another issue, which is permanent contamination of the pumping fluid. Consider
extracting benzene. Enough benzene would dissolve in mineral for it to be considered a hazardous waste. Whether or not you might agree with the degree
of hazard, a large organization would have to consider it as such. Even if you distill off the volatile fraction, the benzene remaining in solution
may still be above permissible levels (I'd have to research this). Getting to below those levels may well require more energy in the distillation than
it's worth. On the other hand, if you're not pumping out contaminants that act that way, this isn't a problem.
One of the advantages of an oil-based ejector would seem to be a greater tolerance to failure of a vapor trap upstream. With a regular rotary vane
vacuum pump, the vapor goes directly into intimate contact with the lubricating oil. This pump is moving gas, so it's tolerances are necessarily
higher. On the other hand, with an ejector pump, the source fluid into the jet and the drain fluid from it can be segregated. The drain fluid can then
be processed (filtered, distilled, etc.) in the course of closing the pumping loop.
Homebrew substitute for a rotary evaporator
watson.fawkes - 14-11-2008 at 09:53
It's occurred to me that there are two fundamental classes of use that vacuum is put to in the chem lab. If you need both of them and only want one
piece of equipment, then use the current practice. On the other hand, it seems that when you have equipment used only for a particular purpose, it may
be more economical to use something else.
The fundamental distinction is thus:
The ultimate pressure achieved, that is, the final result.
The removal of excess material, that is, the means to achieve a result.
For example, if you're evacuating a vacuum line in order to chase volatile reagents from site to site, you care about the resulting pressure. On
the other hand, if you're removing a solvent, you care more about the rate of removal rather than the final pressure (although you might care about
both).
So let's consider the rotary evaporator. It's got rotator, a condenser, and a vacuum hookup. If you use the solvent being evaporated as the pumping
fluid in an ejector pump, the pump also acts as your condenser. You'd need cooling for the pumping fluid, still, but the cooling system can be
different. The potential disadvantage of this method is that the ultimate vapor pressure is higher. But for this application, you'll always have a
vapor pressure gradient because of the difference in temperatures (assuming you run the ejector pump cooler than the evaporator). So maybe the
extraction rate tapers off at the end; that's the cost of using cheaper equipment.
Now if you already own a rotary evaporator (or more; you know who you are) there's no particular need to start using this other method because you're
already invested. On the other hand, if you don't have an adequate means yet, presumably because you can't afford it, then this seems like a
completely workable alternative. It's cheaper, because both the pumping and cooling are easier. And it's much easier to construct.
The big difference with pumps is that you're pumping a liquid instead of a gas, and with the gas, seeking very low pressures to boot. A liquid pump is
always going to be cheaper than a higher-tolerance vacuum pump. It needs to be solvent compatible, but that's pretty easy. Air-powered diaphragm pumps
are particularly useful here, because they're inherently explosion-proof (guarding against vapor release) and specifically designed for swapping out
sealing parts for materials compatibility.
The difference in the cooling system is that you can cool the bulk liquid instead of a vapor. You can use a simple immersion coil, made of copper in
most cases, in the bulk tank holding ejection fluid. You get contact surface area, not through a glass coil, but through the surface area of the spray
out of the ejector nozzle.
All this is easier to construct in the home lab than a rotary evaporator. An ejector pump is mostly some plumbing. The ejector orifice is, at the end
of the day, a hole, albeit an engineered hole. As for cooling, the same pumps and radiators used by PC overclockers should work fine. Or if you need
lower temperatures, the repurposed guts of a mini-fridge would do.
This idea doesn't address the improved boiling characteristics of the rotator system, admittedly.stoichiometric_steve - 14-11-2008 at 12:26
a diaphragm pump (used) is cheap to buy and much more economical than all of those improvised recirculating aspirators - just so you know!watson.fawkes - 14-11-2008 at 14:47
Quote:
Originally posted by stoichiometric_steve
a diaphragm pump (used) is cheap to buy and much more economical than all of those improvised recirculating aspirators - just so you know!
What's the ultimate vacuum they'll pull? I've only seen them pumping liquids.stoichiometric_steve - 14-11-2008 at 23:44
Quote:
Originally posted by watson.fawkesWhat's the ultimate vacuum they'll pull? I've only seen them pumping liquids.
oh, then you should probably read up on the matter. diaphragm pumps are the standard type that is used in labs nowadays when ultimate pressure
>1mmHg is desired.
8-stage pumps go as low as 0.5mmHg.watson.fawkes - 15-11-2008 at 07:04
So I did a bit of poking around about chemically-resistant diaphragm pumps in the market now. Welch Vacuum Products, VacuuBrand Chemical-Resistant Diaphragm Pumps for sale at Cole-Parmer. The cheapest of these, new, is around 1750 USD. At 40 cents on the dollar
for recent surplus, that's still 700 USD. So "cheap to buy" is completely relative here. While I'm sure they're more convenient to use, I fail to see
how they're cheaper for the home lab where, by assumption, there's a different time-vs.-cash trade-off than for a commercial or academic lab.stoichiometric_steve - 15-11-2008 at 12:17
Quote:
Originally posted by watson.fawkes "cheap to buy" is completely relative here.
you probably also know a place called "ebay" where those pumps are regular inventory. i buy them for less than 50USD for a single stage unit and less
than 150USD for a dual stage unit.chemrox - 15-11-2008 at 17:49
I have a two stage mechanical vane pump that pulls 0.1 mm with dicey plumbing attached. I picked up a Welch two stage on LabX that pulls < 1 mm
for $25. The operating costs of these include dry ice and pump oil. I pickjed up an aspirator pump on ebay for $45 and it pulls about 25 mm with
cold tap water. Sometimes I throw in some ice but I also understand solvents are bad for those .. something to do with internal parts and keeping the
motor cool.watson.fawkes - 15-11-2008 at 17:53
So what I'm learning is that these pumps aren't in high demand in the aftermarket. Most of the stuff in the 1-2 K USD range I'm typically looking at
isn't discounted nearly as much as you all are reporting.
