The OP is probably feeding peroxide into a water treatment system.
If you need a flow switch, look at the Efector thermal dispersion switches. They have adjustable switching points and if the lower housing is made with small enough flow passages, they work with flows this low.
Any flowmeter or transmitter you use will need to deal properly with the pulsile output of your pump, and most low-end flowmeter products produce significant errors due to their inability to generate an accurate mean flow out of all those pulses. Use a pulsation dampener to smooth things out so that whatever flowmeter you use will have a better chance of working.
You can forget about the calibration curve method- it won't work with peroxide. The oxygen you need to pump WITH the peroxide will cause you no end of trouble without a flow switch at minimum.
If it's basic indication with switching that you want, once you've got the flow properly pulse-dampened a rotameter will work fine. You can get an optical emitter/detector pair switch which will do a decent job of low flow sensing. But you need to make sure that you vent the oxygen produced by the slow decomposition of the peroxide or it will wreak havoc with your pump's ability to maintain prime- and with the flowmeter's readings. That can be done with a solenoid valve and timer interlocked to the flow switch, or by other means. Some of the pump vendors have their own little tricks to deal with peroxide.
If you actually need accurate flow measurement and transmission at these rates, you're pretty much stuck with either coriolis, thermal (Brooks, Porter, Horiba etc.)or positive displacement flowmeters. If you could find a magmeter with a tiny flowtube, that would work too- 35% is plenty conductive enough- but it's tough to get velocities high enough at these low flows to make the mag principle work properly. None of these options are what you'd call inexpensive, but coriolis is the most expensive by far- and although not totally insensitive to errors induced by any stray 2nd phase (oxygen) in there, it's the method of these which is least sensitive to this problem. Stay away from the laminar flow differential pressure units- they're junk even with water, and will be disastrous with peroxide.
