A logarithmic scale may be more useful. The current could be passed through a diode, and the voltage drop across the diode sensed with a suitable opamp. CMOS opamps are available with bias currents less than 1pA.
It's easy. Refer to the circuit from post #10, calculate a suitable feedback resistor value (500 Meg), change 741 against an "electrometer" OP with sub pA input current (e.g. OPA129) and you're done.
But does a linear, single range I/V converter solve the problem? It possibly does for the restricted 1:2000 range but surely can't for the original 1:10e9 range.
There are at least three possibly solutions:
- multiple ranges with manual selection
- autoranging
- non-linear characteristic like the said log curve
It looks you didn't yet think about a full project specification. I wonder if others (e.g. the project supervisor) did.