Don't touch linux. I hate how people jump right into "Go with Linux" as soon as you are having issues with a Windows machine. There is nothing wrong with Linux, but learning it by installing it as replacement for a production mail server is not the way to learn it.
Unless you are 100% comfortable with Linux, it will just be a bigger pain then its worth and your users have to give up the groupware of Exchange. Since you are non-profit I suspect your users are very non-tech and staying with Exchange/Outlook will be better for them as you said yourself. They also won't like the gmail as much I suspect since the free version (atleast what I've been using for a custom domain account) includes advertising and you need to then move all the email into it. Since you have Exchange and all the licenses already, I'd stay with that and just re-jig your hardware. You and your users will be happier.
Spec wise, 11 users (and the previous 30) will have no issues on either of those. For reference, I ran a 100 user Exchange server, with 40 users active at a time off an older AMD desktop with 512 megs RAM and an IDE drive. This was a temp server while I rebuilt the production one. No one batted an eye for speed, ran great for 4 weeks until I moved it back to the real server.
Here is what I would do,
Spin up the OptiPlex Workstation, setup Windows and Exchange on it.
Move all the mailboxes there, ensure all users are active and referencing the new server (if you leave them both online this will be done automaticly when they log in)
Once that is done, un-install Exchange from the PowerEdge, by un-installing this will remove it from the Exchange Organization cleanly
Remove the PowerEdge from the domain.
Reformat the PowerEdge, make the SCSI drive one large C:
Join it to the domain, reinstall Exchange on it.
Move the mailboxes back over to the PowerEdge server.
Uninstall Exchange from the OptiPlex one everyone is on the new PowerEdge
Remove the OptiPlex from the domain
At this point if the mailboxes are small enough you can leave it like this. But if you can this might be your next steps.
Put the 80 gig IDE into the PowerEdge (I can't remember if the 1400's have an onboard IDE or not, if they do this is easy then)
Format the new 80 gig Drive, and then move the mailbox store to the 80 gig drive.
Once done you now have a 20 gig system drive and an 80 gig data drive.
Setup an automatic backup of the mailbox store to another computer should your data drive fail.
Given the limited money you have and equipment this might be your best way to go.
If you have any $$ at all, add in a RAID 1 card to the 1400 and add 2x 250gig drives, instead of the old SCSI, this would be step 4 of the first part. The rest of the 1400 will last forever, the hard drives are the weak points. The speed and memory of the 1400 will be able to handle what you need. I have a 10 year old HP server that still works like a champ, I dropped a couple of new hard drives in it and its still going strong. Also this was at a not for profit place, so I can relate to the limited budget.
Edit:Long Term For you.
Look at TCO when you are considering what to do. Here is a basic example making assumptions for numbers.
If you take 12 hours to do all the troubleshooting with windows at say $20 per hour that is $240 it will cost your charity. Even if you are an employee you need to count your time as a cost.
Now if you switch to Linux (an OS you don't know so add the learning curve), you need to look at that too. So lets assume you're good and it will take 24 hours to get everything moved to a Linux box and completely online (remember to ExMerge all the mailboxes too). That's about $480 at this point. But also add in any user re-training for the new web client you will be using plus the IMAP setup on Outlook for them. It will likely be more, say add another 1 hour per person x 10 people @ $15 per hours is another $150 you need to factor in. On top of that, add on more time per week for your day to day operations since you will still be getting comfortable with it. All that time you spend on it, is time you don't spead on your other jobs at the charity, hence its a cost.
If you get charity pricing from MS, then say the Cost of License + Your Time is Less then the cost of Learning Linux then it makes more sense to purchase a new License for Windows 2008 (you can use Window 2003 though which will still run on that hardware).
So end of the day, call up CDW get the charity pricing for the Open Licence, then you've got until July 2010 to have the charity save up enough money to purchase that, and in the long run you'll find that much cheaper then going with Linux. I've done the evaluation before, Windows Charity + My time is always cheaper then Linux + Learning Curve + Risk.
I don't know Linux enough (wish I did), so if I got it on a server and it crashed later, I'd be calling in a Linux guru at way more per hour then I make and the charity would take a serious hit (unknown risk). I have yet to encounter a Windows problem I haven't fixed or worst case $260 I can open a ticket with MS (known risk).
And if you do have time, learn Linux, Start it out in a non-critical role so if it crashes on you, you don't need to spend money on someone to get it going again. I am running one linux server that is a backup file server, not the primary one the users use. So if it goes down the primary is still online and I can spend the time figuring it out. Once I'm comfortable with it, I'll likely move it to the production file server so I can free up that Windows license for something else. But until I'm comfortable with it, I can't risk it going down and me not knowing my way around the server. If the IT is a secondary role for you it will be hard for you to spend enough time with Linux on company time, unless you spend time on your learning it. Never a bad idea, but I've been there and I don't like losing that much of my home life to the computer.