opopanax666 - 20-7-2011 at 00:47
Hi everyone,
I'm toying with an idea about identifying plastics, but I'm not sure what I can expect, nor the details of what it would involve. I hope I'm in the
right place to get some info and feedback...
Let me start with the background: I work for a non-profit recycling company, whose main objective is to provide a job to people who are having a hard
time getting work in the regular circuit (ex-detainees, people on long-term health-care, immigrants,... currently about 150 people).
We try as much as possible to recycle the household "stuff" people donate (about 1,500,000 kg per year), but a large part of our waste are plastics.
So we were looking at reducing the amount of waste plastics by sorting them, so they can be recycled. Since people working here are not very
technically proficient, the only thing we can teach them is to check for the recycling codes on the plastics (arrowed triangle & number), but this
only accounts for about 20% of our plastics...
My idea now was to use spectroscopy to check every piece, so that the readout of the device tells them what code it is.
So I called around to check for prices, but at €10,000 to €20,000, this is waaaaay out of our budget (in fact there is no budget to speak of).
So, my question is this: is it possible to diy a spectroscope that could handle the necessary precision to differentiate between the 6 main groups of
plastics (PETE, HDPE, PVC, LDPE, PP & PS), and have some way of letting a computer assess the spectroscopic results (i.e. display the code), so
that the user wouldn't have to?
I found an article about a "DVD spectroscope", and read some posts about using CCD to transfer the data to a computer, so my guess is that it's
possible. I'm "only" looking for a precision of about 50/60%, more being better. And of course as cheap as possible, seeing that until I can provide
some workable results, I'm going to have to finance this project myself...
Any feedback would be greatly appreciated!
matei - 20-7-2011 at 14:03
The only useful (rapid and relatively straightforward) spectroscopic method for polymer identification is infrared spectroscopy. The problem is that
building an IR spectrometer from scratch is just not feasible IMO. The crafty "DVD" spectroscope works in the visible (eventually UV and maybe NIR)
range. Even adapting an old commercial IR spectrometer (think eBay or LabX) for automated identification of plastics won't be easy.
http://goliath.ecnext.com/coms2/gi_0199-498093/Identificatio...
opopanax666 - 21-7-2011 at 08:00
Here in Belgium there are three or four big players in waste recycling, and we work with a few of them. The one we send our plastics to, uses this unit to identify incoming waste plastics (about €20,000 per unit).
Of course the technology involved would be impossible to copy (it's using MEMS, e.g.), but maybe it could be "recreated" as a desktop unit (tungsten
halogen light unit, NIR spectroscope, reflective probe, and a connection to a PC/Labview). It also shows there have to be ways to automatically
interpret the data (certain logarithms maybe), because after scanning, the device simply shows the plastics code on its display.
I'm not saying it would be easy, but possible, no?
[Edited on 22-7-2011 by opopanax666]