Yeah, read the datasheet.... but also don't trust it. It would be nice if the datasheet plotted V/I, because these cells perform very differently than typical solar cells. Whereas you can usually conservatively approximate output power by a percentage of Voc multiplied by a percentage of Isc, performing that approximation using the datasheet claims of Voc and Isc way overshoots even perfect power conversion given a much larger 1cm^2 cell.

I put four of these cells in parallel under roughly one sun of lighting and did not see the claimed Isc using a fluke process calibrator with microAmp resolution.

People -- you need to read the datasheet more closely. These DO NOT produce 50uA at 4V. 50uA is the SHORT CIRCUIT current. At that current the voltage would be in the millivolts. In practice, you'd probably get a few microamps out of them. However, use one of these to charge a supercap and you might be able to run a super-low-power microcontroller for a few ms every couple of minutes. Would be very cool to see what kind of limits you could push with this thing.

(cont)
I think the only real "work" a person is going to get out of this solar cell is enough power to drive a MOSFET, maybe a small 4000-series chip, maybe a PIC or AVR running on a 32Khz crystal. That's about it.
Trickle charge a small battery? Even the smallest LiPo battery sold here, at 20mAh, in a perfect world, with perfect physics, with perfect light, with zero losses in the wiring, with a perfect battery, would take an absolute bare minimum of 400 hours (16 2/3 days) in continuous bright light to charge it up completely (ignoring all the complicated math that goes with charging and discharging).
I think I'll go buy a Shamwow!
JDG

I was thinking about packing several of these on a board, focusing a lens in front of them and polling them to make really crude pictures that I could process with a microcontroller.

People aren't reading the datasheet correctly.
Open circuit maximum voltage output = 4v
Open circuit means zero current flow. Put a load on it and the voltage will drop off dramatically. If you've got maximum voltage, but can't support any current flow, you can't do any work.
Short circuit maximum current output = 50 uA
Again, short circuit means maximum current flow with little or no voltage behind it (hence the term short circuit). If you've got maximum current flow, but no voltage behind it, can't do any work.
What this datasheet doesn't have is a voltage over current graph, which will show the 'knee' of the curve, which would be the optimal region to operate this solar cell.
Ya can't have an open circuit and a short circuit at the same time.
Also, note in the description in the datasheet:
The CPC1822 is a monolithic photovoltaic string of solar cells with switching circuitry. When operating in sunlight or a bright artificial light environment the optical energy will activate the cell array and generate a voltage at the output. The solar cells are capable of generating a floating source voltage and current sufficient to drive and power CMOS ICs, logic gates and/or provide ?trickle charge? for battery applications.
(cont)

does 2 of them have enough juice to charge a phone

Does it have enough power to charge a phone?!?!

Iam making a watch and with a 3.3v ardoino pro mini do u think it would be resonable to add one of these to slowly charge the battery or at least keep it charged for longer?

You could use these micro solar cells as active sensors. Since they generate their own power they would not have to be powered off the controller. I fly quad rotor RC models with lighting on them for show. During the day one of these on a load switch would keep the lights off. In the absence of ambient light the lights would be switched on. Seeing as electric RC models run on batteries, having something that doesn't draw off current would be a plus. There's a board on here somewhere that has opto isolators that gate some mosfet switches. One of these could drive the led in the opto isolator. Hmmm

These would be cool for projects such as heliostats, where you want to track maximum power. You could use a grid of them (say 5) at slightly different angles to work out the angle that is perpendicular to the suns rays...

This minature solar cell worked well for driving a couple of mosfet logic level gates, using only a hi-brightness red led. You don't get much extra power, just enough voltage across a 1 Meg resistor to operate the gate. Previously, i tried the old "red led as a voltage generator" idea, but while that got 2.1 volts in sunshine, it wasn't enough to activate the mosfet with only another red led as light source.

The application note states that it requires a MOSFET for it to work in a recharging circuit. Can anyone (esp Sparkfun) help or link me to a source for this MOSFET or any alternatives as diodes will be too costly a waste of power, according to the app note. Thanks!!!

Can I use this to charge a 3.7v coin cell?
http://www.sparkfun.com/products/10319
My guess would be that the drop from 4v open will result in a voltage that is too low for charging? Then again I could use several of them in series and in parallel to boost voltage and current?

can these be tied to gether to mak eup to a 12 or 18v feed?

It may not "trickle charge" a battery, but it will do something analogous with many microcontrollers. For example, the MSP430G2231 (one of two microcontrollers in the $4.30 TI Launchpad Kit) draws 220uA@2.2V in Active Mode at 1 Mhz, but only 0.5uA in standby. In some monitoring applications, the microcontroller is mostly in standby, but wakes up every second, minute or hour to check a measurement, then goes back to sleep, resulting in a very low average consumption. One of these, charging an ultracap 8 hrs a day might maintain sufficient charge to run the uC as well as sufficient surplus to trigger an alarm signal if the monitored condition goes out of bounds. I've also been working with the Linear Technologies LTC3108, to turn low power sources into usable ones.

Is the datasheet correct in indicating that the internal "diodes" are aligned with the conventional current flow, not against it? If so, wouldn't it be pretty difficult to use these to "trickle charge" a battery?
("Trickle charge" always seems to be in quotes, so I'm wondering if it can really trickle charge anything, or if those words imply something else, like some sort of "simulated" trickle charge.)

4v 50Ua so how many to run an led under room lighting? also maybe use a capacitor and get at least a blink out of an led?

No this will not power a robot but if it produces 50ua in sunlight at 4V that is enough to charge a supercap that can then run some low power micros like the MSP430 from ti or the PIC16F1827.
1 MSP430
1 supercap
1 cpc1822
They all can fit on a 1 inch square pc board and you have a working sun powered micro.

For the price, you could always string together 10 or 20 of these at varying angles :P

Combined with a 0.33F supercap, you can get a full charge (3.9v) in one day. http://dorkbotpdx.org/blog/laen/an_itty_bitty_power_provider
With that power, you can run a 10mA LED for 1 minute. Or power an ATtiny at 32KHz in Idle 24/7 and wake up to take measurements several times a day.
Yes, you need to be very careful about power saving modes and leakage.
No, it will not charge your LiPo minty-boost.

I've seen these used with a super cap to drive an underclocked pic for a couple mins a day. May be useful in a simple outdoor logging application. Something like logging soil moisture or daily sunlight measurements.

There is no switcher, this is a raw solar cell which is effectively a stack of diodes. It will output a voltage relative to the amount of light. Noon-day summer sun I got about 4.2 volts open circuit. Indoors at 5pm indirect light 2 volts. Night indoor lighting 0.4 volts.

Why would you need this? Too little current... and even if you boost the current, will you have enough voltage to drive anything?

So am I correct in thinking that this will give about 4V or nothing. When the light intensity reaches a threshold it'll start outputting 4V but it will stay at zero for any light below that threshold? Is this the "switching circuitry" mentioned?

4 volts from this? I bet it's crazy low current but still, very interesting.