关于Helmholtz分解推导中负号变正号的疑问求助

Hey there, let’s break down that sign confusion you’re hitting in the Helmholtz decomposition— I’ve seen this trip up plenty of people, so you’re definitely not alone. Let’s walk through the key steps where those negative signs turn positive, using the core parts of the derivation you’re looking at.

First: The Scalar Potential Convention #

First off, let’s recall the standard Helmholtz decomposition form:
$$\mathbf{F} = -\nabla\phi + \nabla\times\mathbf{A}$$
The negative sign in front of $\nabla\phi$ is a convention, not a mathematical accident. It’s borrowed from classical field theory (think electrostatic fields: $\mathbf{E} = -\nabla\phi$) to keep consistency with how we define potentials in physics. But this convention is where the first sign puzzle starts.

When we derive $\phi$, we start with the divergence condition:
$$\nabla^2\phi = -\nabla\cdot\mathbf{F}$$
This is a Poisson equation, and to solve it, we use the 3D Green’s function $G(\mathbf{r},\mathbf{r}') = \frac{1}{4\pi|\mathbf{r}-\mathbf{r}'|}$. The integral solution is:
$$\phi(\mathbf{r}) = \frac{1}{4\pi}\int_V \frac{\nabla'\cdot\mathbf{F}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|} dV' + \frac{1}{4\pi}\oint_S \frac{\mathbf{F}(\mathbf{r}')\cdot\hat{\mathbf{n}}'}{|\mathbf{r}-\mathbf{r}'|} dS'$$

The Critical Sign Swap: Derivatives on Source vs. Observation Point #

Here’s where the magic (and confusion) happens. We use integration by parts on the volume integral term. Remember the vector identity:
$$\nabla'\cdot\left(\frac{\mathbf{F}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\right) = \frac{\nabla'\cdot\mathbf{F}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|} + \mathbf{F}(\mathbf{r}')\cdot\nabla'\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right)$$
Rearranging this gives:
$$\frac{\nabla'\cdot\mathbf{F}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|} = \nabla'\cdot\left(\frac{\mathbf{F}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\right) - \mathbf{F}(\mathbf{r}')\cdot\nabla'\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right)$$

Now, here’s the key relation that flips the sign:
$$\nabla'\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right) = -\nabla\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right)$$
Why? Because taking the derivative with respect to the source point $\mathbf{r}'$ is the opposite of taking it with respect to the observation point $\mathbf{r}$. If you move $\mathbf{r}'$ a little in the $x$-direction, it’s the same as moving $\mathbf{r}$ the opposite way— hence the negative sign.

Substitute this into our integral, and the volume integral becomes:
$$\int_V \left[ \nabla'\cdot\left(\frac{\mathbf{F}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\right) + \mathbf{F}(\mathbf{r}')\cdot\nabla\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right) \right] dV'$$
Applying the divergence theorem to the first term turns it into a surface integral (which combines with the original surface integral in $\phi$), leaving us with:
$$\phi(\mathbf{r}) = \text{surface terms} + \frac{1}{4\pi}\int_V \mathbf{F}(\mathbf{r}')\cdot\nabla\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right) dV'$$

Putting It All Together: The Negative Sign Disappears (or Flips) #

Now take the gradient of $\phi$:
$$\nabla\phi = \text{gradient of surface terms} + \frac{1}{4\pi}\nabla\int_V \mathbf{F}(\mathbf{r}')\cdot\nabla\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right) dV'$$
Notice that we can swap the gradient and integral (for well-behaved fields), and $\nabla\left(\nabla\left(\frac{1}{|\mathbf{r}-\mathbf{r}'|}\right)\right)$ is related to the Laplacian. But when we plug back into the decomposition $\mathbf{F} = -\nabla\phi + \nabla\times\mathbf{A}$, the negative sign in front of $\nabla\phi$ cancels with the negative signs from the derivative swaps, resulting in the positive contributions that match the original field $\mathbf{F}$.

For the Vector Potential Term #

The same logic applies to the $\nabla\times\mathbf{A}$ part. We start with $\nabla^2\mathbf{A} = -\nabla\times\mathbf{F}$ (under the Coulomb gauge $\nabla\cdot\mathbf{A}=0$), and when solving via Green’s functions, the derivative swap between $\mathbf{r}$ and $\mathbf{r}'$ again introduces a negative sign that cancels the one in the equation, leading to the positive旋度 term in the decomposition.

In short, the sign flip isn’t a mistake—it’s a combination of a physics convention (the negative sign on the scalar potential) and a mathematical identity (opposite derivatives on source vs. observation points) working together to produce the final form of the decomposition.

内容的提问来源于stack exchange，提问作者Mr Smokey