# Narrowband LED performance compared to Laser for Spectroscopy

I am currently building my own Raman spectrometer and I am in the process of choosing a light source. I am deciding whether I should be using LEDs vs traditional laser modules.

These are my requirements (in order of importance):

1. Price - I am interested to keep the cost of my prototype as low as possible because there is a possibility of this turning into actual mass production. (For the sake of comparison, a "good" LED costs in tens while a "good" laser source can easily fetch in thousands).
2. Performance - Ideally, the output from light source should be predictable and consistent. I understand that LEDs produces a lot of heat and this can cause the LED to output different wavelengths.
3. Energy consumption - I am trying to build a portable unit, so power consumption is a critical aspect of my design.

Traditionally, Raman spectrometers have always employed lasers as their light source. However, I feel this discussion would be helpful to others who are in the same boat as me as there has been significant improvements in LED pricing and performance.

I have also considered using a grating in junction with my LED to focus only a narrow band of light onto my sample but this would result in high energy lost from stray lights.

Note that I am designing this spectrometer to be used exclusively with SERS (Surface enhanced Raman spectroscopy) so my requirements are less stringent as the substrate will greatly enhance the output signal from my sample.

Can someone shed some wisdom on the feasibility of building a low cost light source comprised only of LEDs for my purpose? By the way, I am trying to aim for 785nm.

• Forgive my asking - what about laser pointer devices? They use LED laser, low power (no heat) and even have internal regulation - and they're cheap. – jcoppens Jun 24 '15 at 6:01
• @jcoppens This post on physics SE has some relevant information. physics.stackexchange.com/questions/15757/… – tomnexus Jun 24 '15 at 19:03
• Maybe you can ovenize your light source so that it will remain at a constant temperature. Precision frequency sources sometimes do this. – mkeith Jun 24 '15 at 19:52
• @mkeith Do you have any recommendation on literature covering this subject? I would like to read more – Sparrowcide Jun 25 '15 at 23:50
• Not off the top of my head. I know it is a common practice for frequency sources that need to be very accurate. The basic idea is to keep the actual quartz crystal inside a heated chamber. Temperature sensing and feedback keeps the chamber temperature constant. – mkeith Jun 26 '15 at 1:22

The three requirements you listed are what every designer wants: low price, high performance, low power. My electronics professor used to say that The designer's blanket is always short: if you pull it up to cover your face your feet will be cold, and vice versa.

There are two ways of deciding on this matter:

1. Budget constraint
2. Specification constraint

You have to decide what's more important, then you pick your components.

I'm not an export in Raman spectroscopy, but I assume that there is a relationship between the laser bandwidth and the amount of information you can recover from this technique.

If you decide that cost is paramount you'll be forced to use a laser diode since those are cheaper. If you have a price range, say $100, you can start to shop around and try to find the best laser diode that$100 can buy. When you find what you want, you design your system around the specific diode characteristics.

If you decide that performance is all that matters, you will do the opposite thing. Let's say that you calculated that in order to meet your specifications you need a light source with a bandwidth of 1nm. At this point you can shop around and find the cheaper laser that gives you that bandwidth. Again, at this point you can design your system around it.

This argument obviously applies to every component in your system. It wouldn't make sense to have a very expensive narrow-band laser and then use a cheap monochromator that would kill your information.

If, as you said, price is the number one priority, go with the laser diode.

• Wouldn't LEDs be cheaper than laser diodes since LED technology is more mature and has higher worldwide demand? – Sparrowcide Jun 26 '15 at 20:05
• Probably, but if you're after a narrow-bandwidth light source you want to use a laser diode. See physics.stackexchange.com/questions/173421/… – Leo Jun 26 '15 at 20:08