Just curious if this will provide a more stable voltage with slighty varying loads. Image attached.
It will not divide, both are unity gain 
R2 and R1 don't pass current so the voltage at V- is just Vout.
But more in general, you can indeed make a voltage divider followed by a opamp in unity gain to give you a "voltage divider" in which the load doesn't play a role. But you're still limited in current in what the opamp can supply. Aka, it's not the same as a real voltage regulator.
Like this?
Yes, the opamp will just act as a buffer aka providing power to the output without loading the divider. Because no current will flow into V+ en V- of an opamp.
So for example you can add a 1k to the output of the opamp and the voltage will remain the same. Unlike if you did that with a voltage divider alone because then that 1k would be parallel to R2 an thus mess with the ratio.
But even the opamp is limited in the amount of current it can supply.
not talking about much current. just want to make my circuit at stable as possible. Thank you for the help man! 
Without getting too complicated, the "Voltage Follower" op amp configuration shown in the last post is also
known as a "BUFFER" (Hight Input impedance, Low output impedance). Your use of the word "stable" is
incorrect because the Voltage Follower mimics the input voltage. If the input voltage is "unstable" , the outpout of the Voltage Follower will be identical. What the voltage follower gives you is a way to avoid
loading the voltage divider (because it has a high input impedance). The voltage follower output is rated for the rated output current of the op amp chosen. Op amps are available in power types able to deliver many
times more than ordinary op amps. The first step is stating your design criteria, (ie: the load current for the output of the voltage divider). If the voltage divider is simply a reference for something else, then you need to specify the tolerance criteria for the reference. A voltage divider is dependent on the Vcc used.
So what , exactly are you using the reference voltage for ?
not talking about much current. just want to make my circuit at stable as possible.
You still have not answered the question about what you need the 2.5V for ? (what is the load ?)
If you working with 5V and need 1/2 of it than use a 2.5V voltage regulator
ted:
If you working with 5V and need 1/2 of it than use a 2.5V voltage regulator
Often you are not interested in the exact value but you want 1/2 of supply voltage. In that case it is better to use buffered voltage divider and not voltage regulator/reference. Imagine your circuit is 5V nominal and you use USB power - my computer supply nearly 5.2V (still in specs of USB). Using 2.5V voltage regulator will give you less than desired Vcc/2.
I said - if...,
I am using; 9V + 5V regulator + 2.5V regulator = uniwersal for all aplications, for arduino or not arduino project.
Smajdalf:
- my computer supply nearly 5.2V (still in specs of USB). Using 2.5V voltage regulator will give you less than desired Vcc/2.
That can be calibrated by used code, 0.1V does not make a difference for most projects.
The TL072 opamp requires a minimum of 10V supply to function, its not going to work from 5V single rail.
I think 10V is recommended, sometimes I am using 5V.
The datasheet says clearly the minimum recommended supply is +/-5V for dual rail and the input and output
voltage ranges make it clear it cannot work from 5V single rail at all... These TL072 opamps
are designed to run from +/-15V and were designed 40 years ago. Modern rail-to-rail opamps are much more
useful these days as a +/-15V supply is a rarity now.
That is on papers, I am talking in practice.
If it happens to work for one device for you outside the datasheet parameters, you are lucky, its not the
basis of advice to others. And I suspect the performance(*) is compromised but you didn't notice.
Its not as if much better rail-to-rail opamps aren't available!
(*) bandwidth, current handling, noise figure, gain, distortion, offset voltage, bias current, etc etc.
MarkT:
If it happens to work for one device for you outside the datasheet parameters, you are lucky, its not the
basis of advice to others. And I suspect the performance(*) is compromised but you didn't notice.Its not as if much better rail-to-rail opamps aren't available!
(*) bandwidth, current handling, noise figure, gain, distortion, offset voltage, bias current, etc etc.
Agree.
For 7 kHz and 3MΩ feedback resistor gain = 100.
For noise the best I know is OP27
I would recommend the LT1637
ted:
That is on papers, I am talking in practice.
It is called robust design. If you use a part outside of the specifications, it may work unpredictably and be unreliable. Maybe the temperature goes up or down 5 degrees and it quits working properly.
It is called robust design
Haven't you heard of FOOBAR design ?
The amplifier is a placeholder. probably should have mentioned that but the one i am going to use is burried in a drawer somewhere and i could not be bothered to go find it's name while i threw together the schematic.