Mertens function bound from the classical zeta zero-free region Created 2026-10-03 Updated 2026-10-07
The classical zero-free region of the Riemann zeta function and a truncated Perron formula give
for some absolute constant . Take contour height and move the Perron contour to .
Let converge absolutely at a real , and take . Define
The truncated Perron kernel estimate is
For close the contour to the left, collecting the residue one at zero; for close it to the right, collecting no residue. On the horizontal sides, integrating bounds the error by , and the remote vertical side tends to zero. Near use the bounded transition estimate instead, giving . These are the contours and bounds underlying the kernel formula. The constants are uniform for , the range needed below.
Absolute convergence permits termwise integration. The truncated Perron formula is consequently
Here the primed sum has half weight when is an integer, and otherwise. To obtain the inclusive sum add at an integer. This endpoint convention avoids a false uniform assertion about the kernel at .
Apply this with , , and , for sufficiently large . For , is bounded below, and
using . In the central range , , and . Separate the nearest integers, then sum the harmonic tail over distances : its contribution is . The possible endpoint weight and nearest terms, of size , are absorbed because . Therefore
For , use the same vertical line ; the error estimate remains valid since . Subtract the two formulas. The identity
has absolute value on that line, because is bounded. Taking absolute values gives
This is the short-interval Perron bound for the second Chebyshev function.
The truncated Perron formula says that if converges absolutely for , then for , , and not an integer,
Take and . The logarithmic derivative identity gives . Since is an integer, for every integer . In the range ,
and hence the contribution there is
The ranges and are bounded by the same quantity using absolute convergence and . Therefore
For , and , apply the truncated Perron formula with Von Mangoldt function coefficients at both endpoints on the same vertical line. The difference kernel is , of modulus . The usual near-integer error bound contributes . The logarithmic derivative on this line is finite because of the Euler product.
For , the truncated Perron formula and a contour shift give the displayed form of the Riemann–von Mangoldt explicit formula. The pole at one contributes and the Nontrivial zeros of the Riemann zeta function contribute the finite sum. One can choose a separated height in and then restore using the local zero count for the Riemann zeta function. The error absorbs the half-weight discrepancy at a prime power. Outside this range of , one needs the fuller error with the near-integer terms retained.
For , differs from by . Close the contour to the left for and to the right for ; horizontal sides give the logarithmic denominator and the nearby transition is bounded. At , direct integration gives . This explains the half-weight endpoint in the truncated Perron formula. The constants may be taken uniform for and .