令矩形方框上下边无限接近分界面,那么 l 2 → 0 l_2\to0 l2→0.
⇒ ∮ Γ H ⋅ d l = H 1 ⋅ t l 1 − H 2 ⋅ t l 1 = ( H 1 − H 2 ) ⋅ ( m × n ) l 1 \Rightarrow \oint\limits_\Gamma\mathbf{H}\cdot d\mathbf{l}=\mathbf{H}_1\cdot \mathbf{t}l_1-\mathbf{H_2}\cdot\mathbf{t}l_1=(\mathbf{H}_1-\mathbf{H}_2)\cdot(\mathbf{m}\times\mathbf{n})l_1 ⇒Γ∮H⋅dl=H1⋅tl1−H2⋅tl1=(H1−H2)⋅(m×n)l1
又 I = ∫ Ω J v ⋅ m d s I=\int_\Omega\mathbf{J}_v\cdot\mathbf{m}ds I=∫ΩJv⋅mds,其中 Ω \Omega Ω为矩形线框包围的区域, J v \mathbf{J}_v Jv为体电流密度。由于矩形线框面积趋于0,因此体电流密度产生的电流为0。因此通过矩形线框的只有面电流(可以理解为整个电流压缩在面上),则 I = ∫ L J s ⋅ m d l I=\int\limits_L\mathbf{J}_s\cdot\mathbf{m}dl I=L∫Js⋅mdl
L L L为矩形线框与分界面相交的线段,长度为 l 2 l_2 l2。
因此
( H 1 − H 2 ) ⋅ t l 1 = J s ⋅ m l 1 (\mathbf{H}_1-\mathbf{H}_2)\cdot\mathbf{t}l_1=\mathbf{J}_s\cdot\mathbf{m}l_1 (H1−H2)⋅tl1=Js⋅ml1
⇒ ( H 1 − H 2 ) ⋅ ( m × n ) l 1 = J s ⋅ m l 1 \Rightarrow(\mathbf{H}_1-\mathbf{H}_2)\cdot(\mathbf{m}\times\mathbf{n})l_1=\mathbf{J}_s\cdot\mathbf{m}l_1 ⇒(H1−H2)⋅(m×n)l1=Js⋅ml1
⇒ ( H 1 − H 2 ) ⋅ ( m × n ) = J s ⋅ m \Rightarrow(\mathbf{H}_1-\mathbf{H}_2)\cdot(\mathbf{m}\times\mathbf{n})=\mathbf{J}_s\cdot\mathbf{m} ⇒(H1−H2)⋅(m×n)=Js⋅m
根据混合积的轮换对称性 a ⋅ ( b × c ) = c ⋅ ( a × b ) = b ⋅ ( c × a ) \mathbf{a}\cdot(\mathbf{b}\times\mathbf{c})=\mathbf{c}\cdot(\mathbf{a}\times\mathbf{b})=\mathbf{b}\cdot(\mathbf{c}\times\mathbf{a}) a⋅(b×c)=c⋅(a×b)=b⋅(c×a),
( H 1 − H 2 ) ⋅ ( m × n ) = m ⋅ [ n × ( H 1 − H 2 ) ] (\mathbf{H}_1-\mathbf{H}_2)\cdot(\mathbf{m}\times\mathbf{n})=\mathbf{m}\cdot[\mathbf{n}\times(\mathbf{H}_1-\mathbf{H}_2)] (H1−H2)⋅(m×n)=m⋅[n×(H1−H2)]
所以
m ⋅ [ n × ( H 1 − H 2 ) ] = m ⋅ J s \mathbf{m}\cdot[\mathbf{n}\times(\mathbf{H}_1-\mathbf{H}_2)]=\mathbf{m}\cdot\mathbf{J}_s m⋅[n×(H1−H2)]=m⋅Js
⇒ m ⋅ [ n × ( H 1 − H 2 ) − J s ] = 0 \Rightarrow\mathbf{m}\cdot[\mathbf{n}\times(\mathbf{H}_1-\mathbf{H}_2)-\mathbf{J}_s]=0 ⇒m⋅[n×(H1−H2)−Js]=0
⇒ n × ( H 1 − H 2 ) = J s \Rightarrow\mathbf{n}\times(\mathbf{H}_1-\mathbf{H}_2)=\mathbf{J}_s ⇒n×(H1−H2)=Js
要注意的是 n × H \mathbf{n}\times\mathbf{H} n×H不是 H \mathbf{H} H在 t \mathbf{t} t方向的分量。可以把 H \mathbf{H} H沿着 ( n , t , m ) (\mathbf{n},\mathbf{t},\mathbf{m}) (n,t,m)构成的局部坐标系分解成三个分量, H = H t t + H n n + H m m \mathbf{H}=H_t\mathbf{t}+H_n\mathbf{n}+H_m\mathbf{m} H=Htt+Hnn+Hmm。那么 n × H = H t n × t + H m n × m = H t m − H m t \mathbf{n}\times\mathbf{H}=H_t\mathbf{n}\times\mathbf{t}+H_m\mathbf{n}\times\mathbf{m}=H_t\mathbf{m}-H_m\mathbf{t} n×H=Htn×t+Hmn×m=Htm−Hmt。
如果 J s = 0 \mathbf{J}_s=0 Js=0,那么
n × ( H 1 − H 2 ) = ( H t 1 − H t 2 ) m − ( H m 1 − H m 2 ) t = 0 \mathbf{n}\times(\mathbf{H}_1-\mathbf{H}_2)=(H_{t1}-H_{t2})\mathbf{m}-(H_{m1}-H_{m_2})\mathbf{t}=0 n×(H1−H2)=(Ht1−Ht2)m−(Hm1−Hm2)t=0
因为 t \mathbf{t} t和 m \mathbf{m} m是正交的,因此 H t 1 = H t 2 H_{t1}=H_{t2} Ht1=Ht2, H m 1 = H m 2 H_{m1}=H_{m2} Hm1=Hm2.
上面证明线积分 ∫ Γ H ⋅ d l \int_\Gamma\mathbf{H}\cdot d\mathbf{l} ∫ΓH⋅dl的做法是认为 l 2 = 0 l_2=0 l2=0,所以在上面的线积分为0,这是很多教材上采用的说法。
要让 H \mathbf{H} H在 l 1 l_1 l1上可以视为不变的的, l 1 l_1 l1也应是无穷小。 l 1 l_1 l1和 l 2 l_2 l2都是无穷小,只让 l 2 = 0 l_2=0 l2=0有点说不过去。我觉得沿着矩形线框左右两个侧边的线积分互相抵消等于0,这个解释更合理一点。不过这是个人的观点。
∮ Γ H ⋅ d l = ∫ Γ 左 H d l + ∫ Γ 右 H d l + ∫ Γ 上 H d l + ∫ Γ 下 H d l = ( − H 1 ⋅ n l 2 2 − H 2 ⋅ n l 2 2 ) + ( H 1 ⋅ n l 2 2 + H 2 ⋅ n l 2 2 ) + H 1 ⋅ t l 1 − H 2 ⋅ t l 1 = H 1 ⋅ t l 1 − H 2 ⋅ t l 1 \begin{aligned} \oint\limits_\Gamma\mathbf{H}\cdot d\mathbf{l}=&\int_{\Gamma^{\text{左}}}\mathbf{H}d\mathbf{l}+\int_{\Gamma^{\text{右}}}\mathbf{H}d\mathbf{l}+\int_{\Gamma^{\text{上}}}\mathbf{H}d\mathbf{l}+\int_{\Gamma^{\text{下}}}\mathbf{H}d\mathbf{l}\\ =&(-\mathbf{H}_1\cdot \mathbf{n}\frac{l_2}{2}-\mathbf{H}_2\cdot \mathbf{n}\frac{l_2}{2})+(\mathbf{H}_1\cdot \mathbf{n}\frac{l_2}{2}+\mathbf{H}_2\cdot \mathbf{n}\frac{l_2}{2})+\mathbf{H}_1\cdot \mathbf{t}l_1-\mathbf{H_2}\cdot\mathbf{t}l_1\\ =&\mathbf{H}_1\cdot \mathbf{t}l_1-\mathbf{H_2}\cdot\mathbf{t}l_1 \end{aligned} Γ∮H⋅dl===∫Γ左Hdl+∫Γ右Hdl+∫Γ上Hdl+∫Γ下Hdl(−H1⋅n2l2−H2⋅n2l2)+(H1⋅n2l2+H2⋅n2l2)+H1⋅tl1−H2⋅tl1H1⋅tl1−H2⋅tl1