microfluidics valving question

[bioe007]

I am trying to spec a project, general theory of operation:

have 4 fluids: air(sterile), and 3 aqueous solutions

use the air to flush a (semi) ufluidics card between reagents

the first two reagents are ~mL washes through the entire card (Q=1200uL/min <or less is ok> ), spec grav~ 1.1 , 0.9 - respectively

the final reagent is 100uL and has spec grav=1, it has to flow at about 1/2 the rate of the first two reagents

between each reagent, the card is flushed with the air line.

I am trying to figure out what kind of pumping and valving system is best for this application.

my search so far has turned up:

pump= peristaltic (we use this in the lab, works well)

http://www.ismatec.com/int_e/pumps/t_mcp_bvp/mcp_proc.htm

this pump has a max pressure drop of ~15-22psig (depending on tubing)

valves = micro isolation rocker valves (avsco series 067)

http://www.ascovalve.com/Applications/Products/MicroIsolationData.aspx

tubing something like Tygon MHLL, Norprene

I'd like to run reagents into a block manifold controlled by four of these (or similar valves) then a single line to the pump (with some check valve and/or sterile filter to prevent upstream contamination)

I am new to valving applications and am a bit confounded by figuring out what orifice size I'd need and can only find information about control valves operating at much higher flow rates.

my concerns:

1. how to keep the air sterile, is a .2u filter on the inlet ok?
2. how to compute the valve size needed for each reagent
3. filters, check valves, misc small parts I may not be thinking of?

thanks for looking

 

[MikeHalloran]

It usually makes more sense to run multiple tubes over wide rollers than to share a single peristaltic tube.

Sharing any lumen is not a good idea; at the Reynolds numbers that come with microfluidics, flushing and mixing are both difficult. Bubbles can help with both, but they can be difficult to control... especially if you are still thinking in terms of flow rates. You may do better by thinking in terms of finite volumes. Reagent consumption is a key marketing parameter anyway.

Once upon a time, that 'air flush between cycles' sounded like a good idea. The resulting machine ran okay, so long as you never, ever stopped cycling it. Shut down over a weekend, the damp tubes would precipitate salt crystals, which really did a number on those rocker valves. Other types of valves fared even worse.

I suggest that you minimize the up-front analysis. Sketch up a system, size it based on connection sizes, build it and measure what it does... and do that as fast as possible, because you'll throw it away. Plan on throwing away several more after that, too.







Mike Halloran
Pembroke Pines, FL, USA

 

[bioe007]

thanks for the response,

I suggest that you minimize the up-front analysis. Sketch up a system, size it based on connection sizes, build it and measure what it does...

my real problem is I have no time to prototype this, but need a well designed text description. I really hate engineering this way but sometimes thats how it goes. I certainly know things never work as well at the bench as on paper.

Im getting from your response that perhaps I've got to find an inert flushing reagent, and not use air?

I am thinking in terms of volumes, the only important flow rate is with the last reagent. But the first two I really don't care how fast or slow they move through the card. (sorry, I don't think i presented it that way in my OP)

I should also make it clear there is _no_ mixing that needs to take place in the card, and the volumes involved are >600uL and >100uL. Also the channel geometry is pretty large, and does create turbulent flow.

The reason I want a single peristaltic tube is for scalability. People have already mastered the peristaltic pump, my invention centers around the card and its properties. I want to keep the pump cost minimal and my valving system + cards scalable.

 

[bioe007]

doh, I also forgot, the other constraint for the single line is there is a single input to the card, and that is pretty much frozen.

 

[MikeHalloran]

Miscible flushes, e.g. water or saline, can work in large volumes, or when accompanied by a bubble or two. Generating and removing bubbles adds some complexity, but it has been done reliably.

Immiscible flushes can work too. Some Freons are liquid at ordinary temperatures. I think I have seen them used commercially, but the manufacturer was very tight- lipped about what, exactly, was in the system. It may have been vegetable oil, for all I know.

I would not characterize the peristaltic pump as 'mastered', at least not in systems where precision is important.



Mike Halloran
Pembroke Pines, FL, USA