I remember the time when I was still studying at lukio (equivalent school is high school in the US, gymnasium in Sweden, Gymnasium in Germany) when there were exercises to determine positive integer powers of imaginary unit i. I invented at that time a method to easily determine the power without using the taught method of grouping the exponent. Later on my university times I invented a formula to determine arbitrary real number power of imaginary unit i.
Complex number z is of form z = a + bi, where the previous is the real part and the latter the imaginary part. Now we’re interested only in imaginary unit i. So we examine complex number z = i, the magnitude of i = 1.
The argument (arg) of imaginary unit i is 90°. From this fact it follows that multiplying with i corresponds 90° rotation on the complex plane. Thus, if the exponent is a positive integer, there’s exactly 4 possibilities for the power of i: 1, -1, i, -i.
The power can be determined by dividing the exponent by 4 and examining the decimal part of the division. There are now 4 possibilities: .0; .25; .5; .75. These tell how many percent of unit circle on the complex plane has been rotated.
Based on this the following table can be presented:
The first line is the decimal part of division by 4 of the exponent, the second line tells place in the unit circle of the complex plane, the third line tells the the power of imaginary unit i.
For example to determine i12345678 , we first divide the exponent 12345678 by 4, the result is 3086419.5. The decimal part is .5, so we see from the table above, that the power is -1.
The formula for arbitrary real number exponent
I invented this formula myself some time in the year 1997 when I was studying for the second year at the university of Jyväskylä. I thought my formula was too simple, so I didn’t show it to the maths department staff. Anyway, the formula is as follows:
Many years ago, I asked on local science magazine’s net forum, that does this kind of formula already exist. The existing formula was a bit different. In the same forum one reader presented a proof to this formula.