Are you using the 1K resistor in series with the IR LED? Because based on the datasheet from that page that won't be enough current to really drive it. That will give you about 3.5mA and you need more like 100 mA. So try using a 100 Ohm resistor there instead.
Another cool trick is if you have a camera phone, old webcam or digital camera a lot of times if you point them at your IR LED you'll be able to see it turn on, on screen. Try it yourself by pointing your remote control at your phone and hitting some buttons. Also that might be a good way to see if your detector is outputting anything too. Point it at your IR detector and hold down the volume button and see if you get anything on the Audrino (or even better a scope if you have one).
The wavelength stays more or less the same, but more current = more light or in your case any light at all.
Red might work but I suspect it might be that your LED isn't on.
Good luck!
To emphasise WhatRoughBeast's point, that infra-red and 'daylight' phototransistors are usually the same active, semiconductor, component embedded into plastic with daylight transparent or daylight filtering, these two phototransistors a SFH 300 and a SFH 300 FA have an almost identical specification.
Both types are compatible with an Arduino. A search should find good values for 'voltage divider' resistor.
As Chris Stratton explained, you can take ADC measurements from the same sensor, one illuminated by an IR LED/emitter and the other measurement without IR illumination, then subtract to get the difference.
You can make the illuminated measurement more effective by 'pulsing' a very high current through the IR LED (for a short time). Usually if you look at the spec. sheet for the IR LED, it'll give a value for the maximum pulse current, whicn may be as much as 10x more than the continuous current. You'll need some extra circuitry, for example a bipolar or MOSFET transistor to do this as it'll likely be well beyond the current rating of the Arduino. This might provide enough power to work well using visible light (red) LEDs and sensors.
There are other things to consider with a line follower.
It is possible to get slightly better performance by making sure the LED/emitter and sensor are matched, usually you can get a good match just by getting parts with very similar peak wavelength.
Line follower tracks are often made with a black background and white line, or white with black tape. Black paper, some black paint and some black tape reflects IR light surprisingly well, and looks 'white' in IR light. So check that the track is using materials which have good contrast to your sensors. It is just possible that IR is worse than visible light.
Another thing to consider is using multiple IR sensors and LEDs/emitters. The positions of the IR sensors and IR LED/emitters can help. Search around and you'll find various ideas. If you use multiple sensors and emitters in a reasonable layout, you should get a pretty accurate measurement of the lines position (IIRC, Pete Harrison at http://www.micromouseonline.com/ might have described a technique). A way to think of it is interpolating from a few values, or fitting a curve to a few values. It's worth trying to make a couple of arrangement of sensors and emitters, if you have the time and money.
One small problem with using IR is knowing how well it works. Many mobile phone cameras and digital cameras will display in the viewfinder, and capture IR. You might make good progress by using visible light initially just because it should be easier to see how well things are working.
Best Answer
Like regular transistor a photo-transistor is made with a base, an emitter and a collector but the base is sensible to light allowing more or less current passing through it. So when the base is illuminated your photo-transistor is in saturation state like a regular one.
But in your application I think using a IR diode is not going to be enough. Maybe using a laser pointer is more effective.