静磁場

静磁場(せいじば、英語: static magnetic field)とは、時間的に変動しない磁場のことである [1] [2] [3] [4] [5] [6][7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]

本記事では、静磁気学(Magnetostatics)の視点から、静磁場について述べる。

定常な電流が作り出す静磁場の一般論

本節では、真空中に定常な(つまり時刻tに依存しない)電流密度が作り出す磁束密度について、一般に成り立つ事柄について述べる。ただし、時間的な変動の影響はもちろんのこと、これ以外にも、電場や強制電荷、分極電荷の影響は排除されているものとする。 本記事では、専ら体積電流密度を中心に扱い、線電流近似については、例えば[30] [31] [32]等に委ねることとする。


真空中に定常な(つまり時刻tに依存しない)電流密度 i ( r ) {\displaystyle {\boldsymbol {i}}(\mathbf {r} )} が与えられた とする。このとき、 i ( r ) {\displaystyle {\boldsymbol {i}}(\mathbf {r} )} は、以下の磁気ベクトルポテンシャル A i ( r ) {\displaystyle \mathbf {A} _{\boldsymbol {i}}(\mathbf {r} )} を空間内に作り出す。

A i ( r ) = μ 0 4 π s R 3 ( i ( s ) | r s | )   d 3 s {\displaystyle \mathbf {A} _{\boldsymbol {i}}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \mathbb {R} ^{3}}\left({\frac {{\boldsymbol {i}}(\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\ d^{3}\mathbf {s} } (1-1)

となる。

B i = rot [ A i ] {\displaystyle \mathbf {B} _{\boldsymbol {i}}=\operatorname {rot} [\mathbf {A} _{\boldsymbol {i}}]} を考え併せると、 i ( r ) {\displaystyle {\boldsymbol {i}}(\mathbf {r} )} が直接的に作り出す磁束密度 B i {\displaystyle \mathbf {B} _{\boldsymbol {i}}} は、

B i ( r ) = μ 0 4 π s R 3 ( i ( s ) × ( r s ) | r s | 3 ) d 3 s {\displaystyle \mathbf {B} _{\boldsymbol {i}}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}\int _{\mathbf {s} \in \mathbb {R} ^{3}}\left({\frac {{\boldsymbol {i}}(\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |^{3}}}\right)d^{3}\mathbf {s} } (1-2)

となる。これは、すなわち、ビオ・サバールの法則である。

上記の B i ( r ) {\displaystyle \mathbf {B} _{\boldsymbol {i}}(\mathbf {r} )} に対し、 新たな場 H i {\displaystyle {\mathbf {H} }_{\boldsymbol {i}}} を、

H i ( r ) := 1 μ 0 B i ( r ) {\displaystyle \mathbf {H} _{\boldsymbol {i}}(\mathbf {r} ):={\frac {1}{{\mu }_{0}}}\mathbf {B} _{\boldsymbol {i}}(\mathbf {r} )} (1-3)

と定義する。この場のことを、「電流密度iが作り出す磁場」と呼ぶ。ここでμ0は、真空の透磁率である。 尚、定義の上では、「電流密度iが作り出す磁場」 H i {\displaystyle {\mathbf {H} }_{\boldsymbol {i}}} は、 透磁率がμの場所でも、: H i ( r ) := 1 μ 0 B i ( r ) {\displaystyle \mathbf {H} _{\boldsymbol {i}}(\mathbf {r} ):={\frac {1}{{\mu }_{0}}}\mathbf {B} _{\boldsymbol {i}}(\mathbf {r} )} であることに特に注意されたい。

式(1-2),式(1-3)より、

H i ( r ) = 1 4 π s R 3 ( i ( s ) × ( r s ) | r s | 3 ) d 3 s {\displaystyle \mathbf {H} _{\boldsymbol {i}}(\mathbf {r} )={\frac {1}{4\pi }}\int _{\mathbf {s} \in \mathbb {R} ^{3}}\left({\frac {{\boldsymbol {i}}(\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |^{3}}}\right)d^{3}\mathbf {s} } (1-4)

である。これに、回転微分を作用させると、

rot r [ H i ] = i {\displaystyle \operatorname {rot} _{\mathbf {r} }[\mathbf {H} _{\boldsymbol {i}}]={\boldsymbol {i}}} (1-5)

が得られる。実際、ベクトル解析の公式より、

rot r [ ( i ( s ) × ( r s ) | r s | 3 ) ] = 4 π ( div r [ ( r s ) | r s | 3 ] ) i ( s ) = 4 π δ 3 ( r s ) i ( s ) (1-6) {\displaystyle {\begin{aligned}\operatorname {rot} _{\mathbf {r} }\left[\left({\frac {{\boldsymbol {i}}(\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |^{3}}}\right)\right]&=4\pi \left(\operatorname {div} _{\mathbf {r} }\left[{\frac {(\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |^{3}}}\right]\right){\boldsymbol {i}}(\mathbf {s} )\\&=4\pi {\delta }^{3}(\mathbf {r} -\mathbf {s} ){\boldsymbol {i}}(\mathbf {s} )\qquad {\text{(1-6)}}\end{aligned}}}

従って、

rot r [ H i ( r ) ] = 1 4 π s R 3 rot r [ ( i ( s ) × ( r s ) | r s | 3 ) ] d 3 s = s R 3 δ 3 ( r s ) i ( s ) d 3 s = i ( r ) (1-7) {\displaystyle {\begin{aligned}\operatorname {rot} _{\mathbf {r} }[\mathbf {H} _{\boldsymbol {i}}(\mathbf {r} )]&={\frac {1}{4\pi }}\int _{\mathbf {s} \in \mathbb {R} ^{3}}\operatorname {rot} _{\mathbf {r} }\left[\left({\frac {{\boldsymbol {i}}(\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |^{3}}}\right)\right]d^{3}\mathbf {s} \\&={\int }_{\mathbf {s} \in \mathbb {R} ^{3}}{\delta }^{3}(\mathbf {r} -\mathbf {s} ){\boldsymbol {i}}(\mathbf {s} )d^{3}\mathbf {s} \\&={\boldsymbol {i}}(\mathbf {r} )\qquad {\text{(1-7)}}\end{aligned}}}

が得られる。ここでδ3は、3変数のδ関数(ディラックのデルタ)を意味する。

時間的に定常な磁化が作り出す静磁場の一般論

本節では、定常な(つまり時刻tに依存しない)磁化が作り出す磁束密度について、一般に成り立つ事柄について述べる。ただし、時間的な変動の影響はもちろんのこと、これ以外にも、電場や強制電荷、分極電荷の影響は排除されているものとする。


時間的に定常な磁化が作り出す磁気ベクトルポテンシャル

空間内の領域 Ω {\displaystyle \Omega } に物質が置かれ、前記物質が、定常な(つまり時刻tに依存しない)磁化 M ( r ) {\displaystyle \mathbf {M} (\mathbf {r} )} を帯びているとする。このとき、磁化ベクトルは、「単位体積当たりの磁気モーメントの密度」を表すものであるため、磁化の定義より、

物質内の各点 s Ω {\displaystyle \mathbf {s} \in \Omega } それぞれに、それぞれ
M ( s ) d 3 s {\displaystyle \mathbf {M} (\mathbf {s} ){d}^{3}\mathbf {s} } (2-1-1)
で与えられる磁気モーメントが配置されている

と考えることができる[2]

まず、時間的に定常な磁化が作り出す磁気ベクトルポテンシャルについて考えよう。

(1)原点に置かれた磁気モーメントmは、空間上の位置rに作り出す磁気ベクトルポテンシャルは、

a 0 ( r ) = μ 0 4 π ( m × r | r | ) {\displaystyle \mathbf {a} _{0}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}\left({\frac {\mathbf {m} \times \mathbf {r} }{|\mathbf {r} |}}\right)} (2-1-2)
である。

(2)従って、"(1)"を平行移動すれば、位置 s R 3 {\displaystyle \mathbf {s} \in \mathbb {R} ^{3}} に置かれた磁気モーメントmは、空間上の位置rに作り出す磁気ベクトルポテンシャルは、

a s ( r ) = μ 0 4 π ( m × ( r s ) | r s | ) {\displaystyle \mathbf {a} _{s}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}\left({\frac {\mathbf {m} \times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)} (2-1-3)
である。

従って、上記の磁気モーメント M ( s ) d s {\displaystyle \mathbf {M} (\mathbf {s} )d\mathbf {s} } それぞれは、磁気ベクトルポテンシャル

μ 0 4 π ( ( M ( s ) d s ) × ( r s ) | r s | ) {\displaystyle {\frac {\mu _{0}}{4\pi }}\left({\frac {(\mathbf {M} (\mathbf {s} )d\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)} (2-1-4)

を作り出す。

上記の磁気ベクトルポテンシャルそれぞれを、全ての s Ω {\displaystyle \mathbf {s} \in \Omega } に渡って足し合わせると、

A M ( r ) = μ 0 4 π s Ω ( ( M ( s ) ) × ( r s ) | r s | )   d 3 s {\displaystyle \mathbf {A} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\left({\frac {(\mathbf {M} (\mathbf {s} ))\times (\mathbf {r} -{\mathbf {s} })}{|\mathbf {r} -\mathbf {s} |}}\right)\ {d}^{3}\mathbf {s} } (2-1-5)

を得る。即ち、物質の磁化 M ( r ) {\displaystyle \mathbf {M} (\mathbf {r} )} は、上記の磁気ベクトルポテンシャル A M {\displaystyle \mathbf {A} _{M}} を空間内に作り出す[2] [注釈 1] [注釈 2]

時間的に定常な磁化が作り出す磁束密度

次に、時間的に定常な磁化が作り出す磁束密度について考えよう。磁気ベクトルポテンシャルの回転微分をとれば、磁束密度が得られる。

(1)原点に置かれた磁気モーメントmは、空間上の位置rに作り出す磁気ベクトルポテンシャルは、前述の通り、

a 0 ( r ) = μ 0 4 π ( m × r | r | ) {\displaystyle \mathbf {a} _{0}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}\left({\frac {\mathbf {m} \times \mathbf {r} }{|\mathbf {r} |}}\right)} (2-2-1)
である。これの回転微分をとることで、原点に置かれた磁気モーメントmが、空間上の位置rに作り出す磁束密度は、
b 0 ( r ) = rot r [ μ 0 4 π ( m × r | r | ) ] = μ 0 4 π grad r [ m | r | r | 3 ] + μ 0 4 π div r [ r | r | ] m {\displaystyle \mathbf {b} _{0}(\mathbf {r} )=\operatorname {rot} _{\mathbf {r} }\left[{\frac {\mu _{0}}{4\pi }}\left({\frac {\mathbf {m} \times \mathbf {r} }{|\mathbf {r} |}}\right)\right]={\frac {{\mu }_{0}}{4\pi }}\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {m} |\mathbf {r} \rangle }{{|\mathbf {r} |}^{3}}}\right]+{\frac {{\mu }_{0}}{4\pi }}\operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} }{|\mathbf {r} |}}\right]\mathbf {m} } (2-2-2)
であることが判る。ここで、 <   |   > {\displaystyle <\ |\ >} は、内積を表す。

(2)従って、"(1)"を平行移動すれば、位置 s R 3 {\displaystyle \mathbf {s} \in \mathbb {R} ^{3}} に置かれた磁気モーメントmが、空間上の位置r に作り出す磁束密度が、

b s ( r ) = rot r [ μ 0 4 π ( m × ( r s ) | r s | ) ] = μ 0 4 π   grad r [ m | ( r s ) | r s | 3 ] + μ 0 4 π   div r [ r s | r s | ] m {\displaystyle \mathbf {b} _{s}(\mathbf {r} )=\operatorname {rot} _{\mathbf {r} }\left[{\frac {\mu _{0}}{4\pi }}\left({\frac {\mathbf {m} \times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\right]={\frac {{\mu }_{0}}{4\pi }}\ \operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {m} |(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]+{\frac {{\mu }_{0}}{4\pi }}\ \operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |}}\right]\mathbf {m} } (2-2-3)
であることが判る。

従って、上記の磁気モーメント M ( s ) d s {\displaystyle \mathbf {M} (\mathbf {s} )d\mathbf {s} } それぞれは、磁束密度

μ 0 4 π grad r [ M ( s ) d 3 s | ( r s ) | r s | 3 ] + μ 0 4 π div r [ r s | r s | ] M ( s ) d 3 s {\displaystyle {\frac {{\mu }_{0}}{4\pi }}\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} ){d}^{3}\mathbf {s} |(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]+{\frac {{\mu }_{0}}{4\pi }}\operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |}}\right]\mathbf {M} (\mathbf {s} ){d}^{3}\mathbf {s} } (2-2-4)

を作り出す。

上記の磁気ベクトルポテンシャルそれぞれを、全ての s Ω {\displaystyle \mathbf {s} \in \Omega } に渡って足し合わせると、

B M ( r ) = μ 0 4 π s Ω grad r [ M ( s ) | ( r s ) | r s | 3 ] + div r [ r s | r s | ] M ( s )   d 3 s {\displaystyle \mathbf {B} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]+\operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |}}\right]\mathbf {M} (\mathbf {s} )\ {d}^{3}\mathbf {s} } (2-2-5)

を得る。即ち、物質の磁化 M ( r ) {\displaystyle \mathbf {M} (\mathbf {r} )} は、上記の磁気ベクトルポテンシャル B M {\displaystyle \mathbf {B} _{M}} を空間内に作り出す [2] [注釈 3] [注釈 4]

さて、ベクトル解析の公式から、

M ( s )   =   s Ω   div r [ r s | r s | ] M ( s )   d 3 s {\displaystyle \mathbf {M} (\mathbf {s} )\ =\ {\int }_{\mathbf {s} \in \Omega }\ \operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |}}\right]\mathbf {M} (\mathbf {s} )\ {d}^{3}s} (2-2-6)

が判る。[注釈 5]

従って、上記の磁気モーメント M ( s ) d s {\displaystyle \mathbf {M} (\mathbf {s} )d\mathbf {s} } それぞれが作り出す磁束密度は、

μ 0 4 π grad r [ M ( s ) d 3 s | ( r s ) | r s | 3 ] + μ 0 δ ( r s ) M ( s ) d 3 s {\displaystyle {\frac {{\mu }_{0}}{4\pi }}\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} ){d}^{3}\mathbf {s} |(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]+{\mu }_{0}\delta (\mathbf {r} -\mathbf {s} )\mathbf {M} (\mathbf {s} ){d}^{3}\mathbf {s} } (2-2-7)

と書けることが判る。従って、

B M ( r ) = μ 0 4 π s Ω grad r [ M ( s ) | ( r s ) | r s | 3 ]   d 3 s + μ 0 M ( r ) {\displaystyle \mathbf {B} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]\ {d}^{3}{s}+\mu _{0}\mathbf {M} (\mathbf {r} )} (2-2-8)

を得る。[注釈 6]

今、新たな場 H M {\displaystyle \mathbf {H} _{M}} を、

H M ( r ) := 1 4 π s Ω grad r [ M ( s ) | ( r s ) | r s | 3 ]   d 3 s {\displaystyle \mathbf {H} _{M}(\mathbf {r} ):={\frac {1}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]\ {d}^{3}{s}} (2-2-9)

と定義すると [注釈 7]

B M ( r ) = μ 0 ( H M ( r ) + M ( r ) ) {\displaystyle \mathbf {B} _{M}(\mathbf {r} )={\mu }_{0}(\mathbf {H} _{M}(\mathbf {r} )+\mathbf {M} (\mathbf {r} ))} (2-2-10)

が得られる。

さて、以上の議論から「物質の磁化Mが既知である場合に限ればその磁化Mが作り出す磁束密度BMを計算する術が得られた」 ことになる。然しながら、物質の磁化Mが既知でない場合には、上述の関係式のみからは、BMMも、判らない。上述の関係式は、 一つの拘束条件を与えているに過ぎないのである。(だが、極めて重要な関係式ではある)。

例えば、磁化Miniを帯びた鉄心が空間におかれていたとき、外部からの磁束密度 B ext ( r ) {\displaystyle \mathbf {B} _{\text{ext}}(\mathbf {r} )} が与えられたとき、鉄心の磁化は、元々の磁化Miniと、外部からの磁束密度 B ext ( r ) {\displaystyle \mathbf {B} _{\text{ext}}(\mathbf {r} )} の影響で、元々の磁化Miniとは異なる新たな磁化Mconを得ることになる。仮に、このMcon が判れば、全系の磁束密度が計算できるのだが、

(難所)元々磁化Miniを帯びている物質に、外部から磁束密度 B ext ( r ) {\displaystyle \mathbf {B} _{\text{ext}}(\mathbf {r} )} を印加したとき、最終的に、物質がどのようなMconを得るか?

が、実のところは難しい。そこで、一般には。B-H曲線等の実測結果と、上記の拘束条件を考え合わせ、数値計算によって磁束密度や磁化が計算されるのである。但し、線形物質に関して言えば、上の難所は比較的簡単である。このような特殊な物質に関する問題については、次章で述べることにする。

磁荷密度の導入

空間内の領域 Ω {\displaystyle \Omega } に物質が置かれ、前記物質が、定常な(つまり時刻tに依存しない)磁化 M ( r ) {\displaystyle \mathbf {M} (\mathbf {r} )} を帯びているとする。

先に定義した、(磁化による)磁場

H M ( r ) = 1 4 π s Ω grad r [ M ( s ) | ( r s ) | r s | 3 ]   d 3 s {\displaystyle \mathbf {H} _{M}(\mathbf {r} )={\frac {1}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]\ {d}^{3}{s}} (2-3-1)

の原因が、磁荷であると考えた場合に、それは、どのようなものであるのかを検討しよう。

ベクトル解析の公式より、

div s [ M ( s ) | r s | ] = grad s [ 1 | r s | ] | M ( s ) + div s [ M ] | r s | = r s | r s | 3 | M ( s ) + div s [ M ] | r s | = M ( s ) | ( r s ) | r s | 3 + div s [ M ] | r s | (2-3-2) {\displaystyle {\begin{aligned}\operatorname {div} _{s}\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]&=\left\langle \operatorname {grad} _{\mathbf {s} }\left[{\frac {1}{|\mathbf {r} -\mathbf {s} |}}\right]\left|\right.\mathbf {M} (\mathbf {s} )\right\rangle +{\frac {\operatorname {div} _{\mathbf {s} }[\mathbf {M} ]}{|\mathbf {r} -\mathbf {s} |}}\\&=\left\langle {\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |^{3}}}\left|\right.\mathbf {M} (\mathbf {s} )\right\rangle +{\frac {\operatorname {div} _{\mathbf {s} }[\mathbf {M} ]}{|\mathbf {r} -\mathbf {s} |}}\\&={\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}+{\frac {\operatorname {div} _{\mathbf {s} }[\mathbf {M} ]}{|\mathbf {r} -\mathbf {s} |}}\qquad {\text{(2-3-2)}}\end{aligned}}}

従って、

M ( s ) | ( r s ) | r s | 3 = div s [ M ( s ) | r s | ] div s [ M ] | r s | {\displaystyle {\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}=\operatorname {div} _{s}\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]-{\frac {\operatorname {div} _{\mathbf {s} }[\mathbf {M} ]}{|\mathbf {r} -\mathbf {s} |}}} (2-3-3)

今、スカラー値関数 ϕ M {\displaystyle {\phi }_{M}} を、

ϕ M ( r ) := s Ω M ( s ) | ( r s ) | r s | 3   d 3 Ω {\displaystyle {\phi }_{M}(\mathbf {r} ):=-{\int }_{\mathbf {s} \in \Omega }{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\ {d}^{3}\Omega } (2-3-4)

と定める(磁化が定める磁位、あるいは、磁化が定めるスカラーポテンシャルと言う)と、

ϕ M ( r ) = s Ω div s [ M ( s ) | r s | ]   d 3 Ω + s Ω div s [ M ] | r s |   d 3 Ω {\displaystyle {\phi }_{M}(\mathbf {r} )=-{\int }_{\mathbf {s} \in \Omega }\operatorname {div} _{s}\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\ {d}^{3}\Omega +{\int }_{\mathbf {s} \in \Omega }{\frac {\operatorname {div} _{\mathbf {s} }[\mathbf {M} ]}{|\mathbf {r} -\mathbf {s} |}}\ {d}^{3}\Omega } (2-3-5)

右辺第一項に、ガウスの発散定理を適用すると、

s Ω div s [ M ( s ) | r s | ]   d 3 Ω = s Ω M ( s ) | r s |   d 2 ( Ω ) {\displaystyle {\int }_{\mathbf {s} \in \Omega }\operatorname {div} _{s}\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\ {d}^{3}\Omega ={\int }_{\mathbf {s} \in \partial \Omega }{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\ {d}^{2}(\partial \Omega )} (2-3-6)

さらに、ベクトル解析の公式を適用すると、

s Ω M ( s ) | r s |   d 2 ( Ω ) = s Ω M ( s ) | n Ω | r s |   | d 2 ( Ω ) | {\displaystyle {\int }_{\mathbf {s} \in \partial \Omega }{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\ {d}^{2}(\partial \Omega )={\int }_{\mathbf {s} \in \partial \Omega }{\frac {\langle \mathbf {M} (\mathbf {s} )|\mathbf {n} _{\partial \Omega }\rangle }{|\mathbf {r} -\mathbf {s} |}}\ |{d}^{2}(\partial \Omega )|} (2-3-7)

従って、

ϕ M ( r ) = s Ω div s [ M ( s ) | r s | ]   d 3 Ω   + s Ω M ( s ) | n Ω | r s |   | d 2 ( Ω ) | {\displaystyle {\phi }_{M}(\mathbf {r} )=-{\int }_{\mathbf {s} \in \Omega }\operatorname {div} _{s}\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\ {d}^{3}\Omega \ +{\int }_{\mathbf {s} \in \partial \Omega }{\frac {\langle \mathbf {M} (\mathbf {s} )|\mathbf {n} _{\partial \Omega }\rangle }{|\mathbf {r} -\mathbf {s} |}}\ |{d}^{2}(\partial \Omega )|} (2-3-8)

ここで、 Ω {\displaystyle \partial \Omega } は、領域 Ω {\displaystyle \Omega } の境界面を意味する。 また、 n Ω {\displaystyle \mathbf {n} _{\partial \Omega }} は、 Ω {\displaystyle \partial \Omega } の法線ベクトルを意味する。

今、体積磁荷密度と、表面磁荷密度を、

  • ρ M := div r [ M ] {\displaystyle {\rho }_{M}:=-\operatorname {div} _{\mathbf {r} }[\mathbf {M} ]} (体積磁荷密度) (2-3-10)
  • σ M :=< M   |   n Ω > {\displaystyle {\sigma }_{M}:=<\mathbf {M} \ |\ \mathbf {n} _{\partial \Omega }>} (表面磁荷密度) (2-3-11)

により定めると、

ϕ M ( r ) = s Ω ρ M ( s ) | r s |   d 3 Ω   + s Ω σ M ( s ) | r s |   | d 2 ( Ω ) | {\displaystyle {\phi }_{M}(\mathbf {r} )={\int }_{\mathbf {s} \in \Omega }{\frac {{\rho }_{M}(\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\ {d}^{3}\Omega \ +{\int }_{\mathbf {s} \in \partial \Omega }{\frac {{\sigma }_{M}(\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\ |{d}^{2}(\partial \Omega )|} (2-3-12)

を得る。

一方、 ϕ M ( r ) {\displaystyle {\phi }_{M}(\mathbf {r} )} の定義により、

H M ( r ) = grad r [ ϕ M ( r ) ] {\displaystyle {\mathbf {H} }_{M}(\mathbf {r} )=-\operatorname {grad} _{\mathbf {r} }[{\phi }_{M}(\mathbf {r} )]} (2-3-13)

であるため、

H M ( r ) = s Ω grad r [ ρ M ( s ) | r s | ]   d 3 Ω   + s Ω grad r [ σ M ( s ) | r s | ]   | d 2 ( Ω ) | {\displaystyle {\mathbf {H} }_{M}(\mathbf {r} )={\int }_{\mathbf {s} \in \Omega }\operatorname {grad} _{\mathbf {r} }\left[{\frac {{-\rho }_{M}(\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\ {d}^{3}\Omega \ +{\int }_{\mathbf {s} \in \partial \Omega }\operatorname {grad} _{\mathbf {r} }\left[{\frac {{-\sigma }_{M}(\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\ |{d}^{2}(\partial \Omega )|} (2-3-14)

である。

磁化電流の導入

時間的に定常な磁化が作り出す磁気ベクトルポテンシャルを、別の側面から考察してみることにしよう。ここでは、「磁束密度の原因は、電流に帰される」という思想に従い、 だとすれば、「磁化と等価な効果を発揮する電流」がどのようなものかを検討することにする。

前記 A M ( r ) {\displaystyle \mathbf {A} _{M}(\mathbf {r} )} に、 ベクトル解析の公式を適用すると、

A M ( r ) = μ 0 4 π s Ω ( ( M ( s ) ) × ( r s ) | r s | )   d 3 s = μ 0 4 π s Ω ( rot s [ M ( s ) ] | r s | )   d 3 s + μ 0 4 π s Ω ( M ( s ) × n Ω | r s | )   | d 2 ( Ω ) | (2-4-1) {\displaystyle {\begin{aligned}\mathbf {A} _{M}(\mathbf {r} )&={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\left({\frac {(\mathbf {M} (\mathbf {s} ))\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\ {d}^{3}\mathbf {s} \\&={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\left({\frac {\operatorname {rot} _{\mathbf {s} }[\mathbf {M} (\mathbf {s} )]}{|\mathbf {r} -\mathbf {s} |}}\right)\ {d}^{3}\mathbf {s} +{\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \partial \Omega }\left({\frac {\mathbf {M} (\mathbf {s} )\times \mathbf {n} _{\partial \Omega }}{|\mathbf {r} -\mathbf {s} |}}\right)\ |{d}^{2}(\partial \Omega )|\qquad {\text{(2-4-1)}}\end{aligned}}}

となることが判る。ここで、 rot s {\displaystyle \operatorname {rot} _{\mathbf {s} }} は、変数 s {\displaystyle \mathbf {s} } についての回転微分を意味する。 Ω {\displaystyle \partial \Omega } は、領域 Ω {\displaystyle \Omega } の境界を意味する。 上式の右辺第二項の積分において、「 Ω {\displaystyle \partial \Omega } の面素の絶対値」を意味し 所謂普通の面積分ではない(この「面積分」は、「ベクトル場からベクトルを作る」特殊なもの)ので注意が必要である。 この積分において、 n Ω {\displaystyle \mathbf {n} _{\partial \Omega }} は、 Ω {\displaystyle \partial \Omega } の法線ベクトルを意味する。

実際、スカラー倍の回転微分の公式より、

rot s [ M ( s ) | r s | ] = ( grad s [ M ( s ) | r s | ] ) × M ( s ) + rot s [ M ( s ) ] | r s | = ( r s ) × M ( s ) | r s | 3 + rot s [ M ( s ) ] | r s | (2-4-2) {\displaystyle {\begin{aligned}\operatorname {rot} _{\mathbf {s} }\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]&=\left(\operatorname {grad} _{\mathbf {s} }\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\right)\times \mathbf {M} (\mathbf {s} )+{\frac {\operatorname {rot} _{\mathbf {s} }\left[\mathbf {M} (\mathbf {s} )\right]}{|\mathbf {r} -\mathbf {s} |}}\\&={\frac {(\mathbf {r} -\mathbf {s} )\times \mathbf {M} (\mathbf {s} )}{{|\mathbf {r} -\mathbf {s} |}^{3}}}+{\frac {\operatorname {rot} _{\mathbf {s} }\left[\mathbf {M} (\mathbf {s} )\right]}{|\mathbf {r} -\mathbf {s} |}}\qquad {\text{(2-4-2)}}\end{aligned}}}

従って、

M ( s ) × ( r s ) | r s | 3 = rot s [ M ( s ) ] | r s | rot s [ M ( s ) | r s | ] {\displaystyle {\frac {\mathbf {M} (\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{{|\mathbf {r} -\mathbf {s} |}^{3}}}={\frac {\operatorname {rot} _{\mathbf {s} }\left[\mathbf {M} (\mathbf {s} )\right]}{|\mathbf {r} -\mathbf {s} |}}-\operatorname {rot} _{\mathbf {s} }\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]} (2-4-3)

従って、

A M ( r ) = μ 0 4 π s Ω rot s [ M ( s ) ] | r s |   d 3 s   μ 0 4 π s Ω rot s [ M ( s ) | r s | ]   d 3 s   {\displaystyle \mathbf {A} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }{\frac {\operatorname {rot} _{\mathbf {s} }\left[\mathbf {M} (\mathbf {s} )\right]}{|\mathbf {r} -\mathbf {s} |}}\ {d}^{3}\mathbf {s} \ -{\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {rot} _{\mathbf {s} }\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\ {d}^{3}\mathbf {s} \ }   (2-4-4)

を得る。さらに、上式の右辺第二項の積分にベクトル解析の公式を適用すると、

μ 0 4 π s Ω rot s [ M ( s ) | r s | ]   d 3 s   = μ 0 4 π s Ω ( M ( s ) | r s | ) × d 2 s   {\displaystyle -{\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {rot} _{\mathbf {s} }\left[{\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right]\ {d}^{3}\mathbf {s} \ ={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \partial \Omega }\left({\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\times {d}^{2}\mathbf {s} \ }   (2-4-5)

を得る。さらに、右辺にベクトル解析の公式を適用すると、

μ 0 4 π s Ω ( M ( s ) | r s | ) × d 2 s   = μ 0 4 π s Ω ( M ( s ) × n Ω | r s | )   | d 2 ( Ω ) |   {\displaystyle {\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \partial \Omega }\left({\frac {\mathbf {M} (\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\times {d}^{2}\mathbf {s} \ ={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \partial \Omega }\left({\frac {\mathbf {M} (\mathbf {s} )\times \mathbf {n} _{\partial \Omega }}{|\mathbf {r} -\mathbf {s} |}}\right)\ |{d}^{2}(\partial \Omega )|\ } (2-4-6)

が得られる。以上から示すべき式が証明された。

今、 i M ( r ) ,   K M {\displaystyle {\boldsymbol {i}}_{M}(\mathbf {r} ),\ {\mathbf {K} }_{M}} を、

  • i M ( r ) := rot [ M ( r ) ] {\displaystyle {\boldsymbol {i}}_{M}(\mathbf {r} ):=\operatorname {rot} [\mathbf {M} (\mathbf {r} )]} (体積磁化電流密度) (2-4-7)
  • K M ( r ) := M ( r ) × n Ω ( r ) {\displaystyle {\mathbf {K} }_{M}(\mathbf {r} ):=\mathbf {M} (\mathbf {r} )\times \mathbf {n} _{\partial \Omega }(\mathbf {r} )} (表面磁化電流密度) (2-4-8)

と置くと、結局、

A M ( r ) = μ 0 4 π s Ω ( i M ( s ) | r s | )   d 3 s +   μ 0 4 π s Ω ( K M ( s ) | r s | )   | d 2 ( Ω ) |   {\displaystyle \mathbf {A} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\left({\frac {{\boldsymbol {i}}_{M}(\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\ {d}^{3}\mathbf {s} +\ {\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \partial \Omega }\left({\frac {\mathbf {K} _{M}(\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\ |{d}^{2}(\partial \Omega )|\ } (2-4-9)

となる。

両辺の回転微分を取ると、

B M ( r ) = μ 0 4 π s R 3 ( i M ( s ) × ( r s ) | r s | 3 ) d 3 s   +   μ 0 4 π s Ω ( K M ( s ) × ( r s ) | r s | 3 )   | d 2 ( Ω ) | {\displaystyle \mathbf {B} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}\int _{\mathbf {s} \in \mathbb {R} ^{3}}\left({\frac {{\boldsymbol {i}}_{M}(\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |^{3}}}\right)d^{3}\mathbf {s} \ +\ {\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \partial \Omega }\left({\frac {\mathbf {K} _{M}(\mathbf {s} )\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |^{3}}}\right)\ |{d}^{2}(\partial \Omega )|} (2-4-10)

が、判る。この見方は、特に、磁化が一様な場合(より一般にはrot[M]=0)といった特殊な場合に、特に威力を発揮する。


強制電流と磁化の両方が既知のとき(一般論)

全系の磁束密度

強制電流ifc (forced current)と磁化Mの両方が既知とし、これら以外に磁束密度の原因となるものがないとした場合、 全系の磁束密度Btot は、強制電流ifcに起因する磁束密度の成分(式(1-2)から求められる) Bfcと、磁化Mが作り出す磁束密度の成分(式(2-2-8)から求められる) BMによって、

Btot=Bfc+BM (3-1-1)

と表される。

ここで、よく注意しておかないといけないことは、「強制電流ifcと磁化Mの両方が既知」という言葉の意味であるが、 仮に、強制電流ifc(や、他の磁化)がない状態での磁化Miniが既知したとして、 (磁化Miniが作り出す磁束密度BMini がどうなるかは(2-2-8)から計算できるが)

Btot=Bfc+BMini (間違った式)

は、よほど特殊な場合を除き成り立たない。要は間違いである。

式(3-1-1)より正確に書くならば、「外場の影響等により磁化が変化した後の磁化 Mcon が作り出す磁束密度BMconを用いて、

Btot=Bfc+BMcon (正しい式)

である。結局以下の難所は残ったままである。

(難所)元々磁化Miniを帯びている物質に、外部から磁束密度 B ext ( r ) {\displaystyle \mathbf {B} _{\text{ext}}(\mathbf {r} )} を印加したとき、最終的に、物質がどのようなMconを得るか?

この問題が、実のところは難しい。つまり、強制電流が作った磁束密度や、磁化自身が作り出す磁束密度により、物質の磁化が最初の磁化から変化してしまうという問題がある。 一般には。B-H曲線等の実測結果と、上記の拘束条件を考え合わせ、数値計算によって磁束密度や磁化が計算されるのである。但し、線形物質に関して言えば、上の難所は比較的簡単である。このような特殊な物質に関する問題については、次章で述べることにする。

再び式(3-1-1)について考えよう。ベクトル解析の公式から

div [ B f c ] = 0 {\displaystyle \operatorname {div} [\mathbf {B} _{fc}]=0} (3-1-2a)
div [ B M ] = 0 {\displaystyle \operatorname {div} [\mathbf {B} _{M}]=0} (3-1-2b)
div [ B tot ] = 0 {\displaystyle \operatorname {div} [\mathbf {B} _{\text{tot}}]=0} (3-1-2c)

であることが判る。即ち、磁束保存の式が満たされることが判る。

B=μ0(H+M)について

前節同様に、強制電流ifcと磁化Mの両方が既知とし、これら以外に磁束密度の原因となるものがないとした場合 について考える。式(3-1-1)のBtotに対し、新たな場Htotを、

H tot ( r ) := B tot ( r ) μ 0 M ( r ) {\displaystyle \mathbf {H} _{\text{tot}}(\mathbf {r} ):={\frac {\mathbf {B} _{\text{tot}}(\mathbf {r} )}{{\mu }_{0}}}-\mathbf {M} (\mathbf {r} )} (3-2-1)

と定める。式(3-2-1)は、所謂”B=μ0(H+M)”に他ならない。


さらに、強制電流 i f c {\displaystyle {\boldsymbol {i}}_{fc}} が作る磁束密度 B f c ( r ) {\displaystyle \mathbf {B} _{fc}(\mathbf {r} )} は、式(1-2)に より、(周辺の物質の有無にかかわらず)定まるが、式(1-3)に倣い 「強制電流 i f c {\displaystyle {\boldsymbol {i}}_{fc}} が作る磁場 H f c ( r ) {\displaystyle \mathbf {H} _{fc}(\mathbf {r} )} 」を、

H f c ( r ) := 1 μ 0 B f c ( r ) {\displaystyle \mathbf {H} _{fc}(\mathbf {r} ):={\frac {1}{{\mu }_{0}}}\mathbf {B} _{fc}(\mathbf {r} )} (3-2-2)

によって定め、磁化による磁場HMを式(2-2-10)のように定めると、 式(3-2-1)と式(3-2-2)、式(2-2-10)より、

H tot ( r ) = B tot ( r ) μ 0 M ( r ) = B f c ( r ) + B M ( r ) μ 0 M ( r ) = H f c ( r ) + ( H M ( r ) + M ( r ) ) M ( r ) = H f c ( r ) + H M ( r ) (3-2-3) {\displaystyle {\begin{aligned}\mathbf {H} _{\text{tot}}(\mathbf {r} )&={\frac {\mathbf {B} _{\text{tot}}(\mathbf {r} )}{{\mu }_{0}}}-\mathbf {M} (\mathbf {r} )\\&={\frac {\mathbf {B} _{fc}(\mathbf {r} )+\mathbf {B} _{M}(\mathbf {r} )}{{\mu }_{0}}}-\mathbf {M} (\mathbf {r} )\\&=\mathbf {H} _{fc}(\mathbf {r} )+(\mathbf {H} _{M}(\mathbf {r} )+\mathbf {M} (\mathbf {r} ))-\mathbf {M} (\mathbf {r} )\\&=\mathbf {H} _{fc}(\mathbf {r} )+\mathbf {H} _{M}(\mathbf {r} )\qquad {\text{(3-2-3)}}\end{aligned}}}

が得られる。従って、

rot [ H tot ( r ) ] = rot [ H f c ( r ) ] + rot [ H M ( r ) ] = i f c {\displaystyle \operatorname {rot} [\mathbf {H} _{\text{tot}}(\mathbf {r} )]=\operatorname {rot} [\mathbf {H} _{fc}(\mathbf {r} )]+\operatorname {rot} [\mathbf {H} _{M}(\mathbf {r} )]={\boldsymbol {i}}_{fc}} (3-2-4)

が得られる。これは、即ち(静磁場の)アンペールの法則である。

式(3-2-4)を示そう。式(2-2-9) 及び以下の式(3-2-5)より、

rot grad=0 (3-2-5)

H M ( r ) {\displaystyle \mathbf {H} _{M}(\mathbf {r} )} に対し、回転微分を作用させると、

rot [ H M ( r ) ] = 1 4 π s Ω rot r [ grad r [ M ( s ) | ( r s ) | r s | 3 ] ]   d 3 s   = 0 {\displaystyle \operatorname {rot} [\mathbf {H} _{M}(\mathbf {r} )]={\frac {1}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {rot} _{\mathbf {r} }\left[\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]\right]\ {d}^{3}{s}\ =0} (3-2-6)

が判る。従って、磁化に起因する磁場 H M ( r ) {\displaystyle \mathbf {H} _{M}(\mathbf {r} )} は、

rot [ H M ( r ) ] = 0 {\displaystyle \operatorname {rot} [\mathbf {H} _{M}(\mathbf {r} )]=0} (3-2-7)

を満たす。


式(3-2-7)と式(1-5)より、式(3-2-4)が示された。

透磁率の導入

透磁率なる概念を、新たに導入する。 即ち、3次正方行列に値を取る行列値関数 μ ( r ) {\displaystyle {\mu }(\mathbf {r} )} (透磁率)を用いて、

B tot ( r ) = μ ( r ) H tot ( r ) {\displaystyle \mathbf {B} _{\text{tot}}(\mathbf {r} )={\mu }(\mathbf {r} )\mathbf {H} _{\text{tot}}(\mathbf {r} )} (3-3-1)

と書けるものとする。この透磁率は、磁化の概念の”すり替え”に過ぎない概念である。 尚、通常は、透磁率は、スカラー値関数と考えてよい場合が多いのだが、その場合は、 「対角成分が全部同じ値で、それ以外の成分が0の行列値関数と、スカラー値関数が同一視できる」ことを思い起こせばいい。

さらに、式(3-3-1) で見たような、透磁率(透磁率テンソル)を用いて、

B tot ( r ) = μ ( r ) H tot ( r ) {\displaystyle \mathbf {B} _{\text{tot}}(\mathbf {r} )={\mu }(\mathbf {r} )\mathbf {H} _{\text{tot}}(\mathbf {r} )} (3-3-2)

と書き表せ、かつ、μが全点で正則行列(逆行列を持つ)とする。(そう考えても、”あまり”一般性を失わない)逆に言えば、

H tot ( r ) = ν ( r ) B tot ( r ) {\displaystyle \mathbf {H} _{\text{tot}}(\mathbf {r} )={\nu }(\mathbf {r} )\mathbf {B} _{\text{tot}}(\mathbf {r} )} (3-3-3)

である。ここで、

ν:=μ-1 (3-3-4)

は、各点でμの逆行列を与えるような行列値関数である。即ち、

μ(s) ν(s)=μ(s) μ-1(s)=E3 (3-3-5)

を充たすような行列値関数である。(ここでE3は3次の単位行列)である。


これまでの議論では、電流素片や磁気双極子モーメントが作り出した磁束密度/磁場について論じてきたが、 式(3-1-2),式(3-2-4)より、このようにして作られた全系の磁場が、磁束保存の式と、アンペールの法則を 充たすことが判った。 ここからは、逆に、磁束保存の式(3-3-6)と、アンペールの法則(3-3-7)、即ち、

div [ B f c ] = 0 {\displaystyle \operatorname {div} [\mathbf {B} _{fc}]=0} (3-3-6a)
div [ B M ] = 0 {\displaystyle \operatorname {div} [\mathbf {B} _{M}]=0} (3-3-6b)
div [ B tot ] = 0 {\displaystyle \operatorname {div} [\mathbf {B} _{\text{tot}}]=0} (3-3-6c)
rot [ H tot ] = i f c {\displaystyle \operatorname {rot} [\mathbf {H} _{\text{tot}}]={\boldsymbol {i}}_{fc}} (3-3-7)

を出発点とし、全系の磁気ベクトルポテンシャル A tot {\displaystyle \mathbf {A} _{\text{tot}}} が充たす微分方程式を導出する。

まず、磁束保存の式 (3-3-6c)は、強制電流に起因する成分 (3-3-6a)、磁化に起因する成分 (3-3-6b) それぞれについて成り立つため、(ポアンカレの補助定理より)それぞれがベクトルポテンシャルを持ち、

rot [ A f c ] = B f c {\displaystyle \operatorname {rot} [\mathbf {A} _{fc}]=\mathbf {B} _{fc}} (3-3-8a)
rot [ A M ] = B M {\displaystyle \operatorname {rot} [\mathbf {A} _{M}]=\mathbf {B} _{M}} (3-3-8b)

をみたすようなベクトル場 A f c {\displaystyle \mathbf {A} _{fc}} と、 A M {\displaystyle \mathbf {A} _{M}} が存在する。 このようなベクトル場 A f c {\displaystyle \mathbf {A} _{fc}} と、 A M {\displaystyle \mathbf {A} _{M}} は、ゲージ不定性を除き一意に定まるが、 本記事では、強制電流に起因する成分 A f c {\displaystyle \mathbf {A} _{fc}} は、式(1-1)のものを採用し、 と、磁化に起因する成分 A M {\displaystyle \mathbf {A} _{M}} 式(2-1-5)を採用することにする。

全系のベクトルポテンシャルを、

A tot := A M + A f c {\displaystyle \mathbf {A} _{\text{tot}}:=\mathbf {A} _{M}+\mathbf {A} _{fc}} (3-3-9)

と定めると、

rot [ A tot ] = B tot {\displaystyle \operatorname {rot} [\mathbf {A} _{\text{tot}}]=\mathbf {B} _{\text{tot}}} (3-3-8c)

が成り立つ。これは、(3-3-6c)の、磁束保存の式をベクトルポテンシャルを用いて書いたものに他ならない。

さらに、式(3-3-8c)に、式 (3-3-2)と、アンペールの法則(3-3-7)を考え併せると、

rot [ ν rot [ A ] ] i f c = 0 {\displaystyle \operatorname {rot} [{\nu }\operatorname {rot} [\mathbf {A} ]]-{\boldsymbol {i}}_{fc}=0} (3-3-10)

が得られ、未知のベクトル場BやHが消え、未知のベクトル場はAのみとなる。これを 「ベクトルポテンシャルによる静磁場の方程式」(静磁場の支配方程式)という。

静磁場の境界条件

物質の境界において、式(3-3-6a)、式 (3-3-7)に、ガウスの発散定理や、ケルビンストークスの定理を適用すると、 磁場および磁束密度の境界条件が得られる。


静磁場のエネルギーと停留値関数

前節の前提条件において、 全系の磁気ベクトルポテンシャル A tot {\displaystyle \mathbf {A} _{\text{tot}}} は、以下の汎関数Fの停留関数となることが判る。[4][8]

F [ A ] := s R 3 1 2 <   H ( s )   |   B ( s ) > <   i f c ( s ) |   A tot ( s ) >   d s = s R 3 1 2 <   ν ( s ) ( rot s [ A ( s ) ] )   |   ( rot s [ A ( s ) ] ) > <   i f c ( s ) |   A ( s ) >   d s (3-5-1) {\displaystyle {\begin{aligned}F[\mathbf {A} ]&:={\int }_{\mathbf {s} \in \mathbb {R} ^{3}}{\frac {1}{2}}<\ \mathbf {H} (\mathbf {s} )\ |\ \mathbf {B} (\mathbf {s} )>-<\ {\boldsymbol {i}}_{fc}(\mathbf {s} )|\ \mathbf {A} _{\text{tot}}(\mathbf {s} )>\ d\mathbf {s} \\&={\int }_{\mathbf {s} \in \mathbb {R} ^{3}}{\frac {1}{2}}<\ {\nu }(\mathbf {s} )(\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )])\ |\ (\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )])>-<\ {\boldsymbol {i}}_{fc}(\mathbf {s} )|\ \mathbf {A} (\mathbf {s} )>\ d\mathbf {s} \qquad {\text{(3-5-1)}}\end{aligned}}}

次元解析をすると、式(3-5-1)は、エネルギーの次元を持つことが判るが、実際に、式(3-5-1)は、全系のエネルギーとなっている。 ここで、 <   |   > {\displaystyle <\ |\ >} は、内積を表す。νの定義は、式(3-3-4)に記載のとおりである。


式(3-5-1)が、本当に、「磁気ベクトルポテンシャルの停留汎関数」ことを検証しよう。 Aに対し、微小な摂動δA[注釈 8] を与えた際の第一変分δF[注釈 8]、即ち、

δF=F[A+δA]-F[A] (3-5-2)

を求める。

F[A+δA]は、以下の被積分関数を、全空間でsについて積分したものである。

1 2 <   ν ( s ) ( rot s [ A ( s ) + δ A ( s ) ] )   |   ( rot s [ A ( s ) + δ A ( s ) ] ) >   <   i f c ( s ) |   A ( s ) + δ A ( s ) > {\displaystyle {\frac {1}{2}}<\ {\nu }(\mathbf {s} )(\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )+\delta \mathbf {A} (\mathbf {s} )])\ |\ (\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )+\delta \mathbf {A} (\mathbf {s} )])>\ -<\ {\boldsymbol {i}}_{fc}(\mathbf {s} )|\ \mathbf {A} (\mathbf {s} )+\delta \mathbf {A} (\mathbf {s} )>} (3-5-3)

一方、F[A]は、以下の被積分関数を、全空間でsについて積分したものである。

1 2 <   ν ( s ) ( rot s [ A ( s ) ] )   |   ( rot s [ A ( s ) ] ) >   <   i f c ( s ) |   A ( s ) > {\displaystyle {\frac {1}{2}}<\ {\nu }(\mathbf {s} )(\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )])\ |\ (\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )])>\ -<\ {\boldsymbol {i}}_{fc}(\mathbf {s} )|\ \mathbf {A} (\mathbf {s} )>} (3-5-4)

したがって、δFは、以下の(3-5-5)を、全空間でsについて積分したものである。

<   ν ( s ) ( rot s [ A ( s ) ] )   |   ( rot s [ δ A ( s ) ] ) > + 1 2 <   ν ( s ) ( rot s [ δ A ( s ) ] )   |   ( rot s [ δ A ( s ) ] ) >   <   i f c ( s ) |   δ A ( s ) > {\displaystyle <\ {\nu }(\mathbf {s} )(\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )])\ |\ (\operatorname {rot} _{\mathbf {s} }[\delta \mathbf {A} (\mathbf {s} )])>+{\frac {1}{2}}<\ {\nu }(\mathbf {s} )(\operatorname {rot} _{\mathbf {s} }[\delta \mathbf {A} (\mathbf {s} )])\ |\ (\operatorname {rot} _{\mathbf {s} }[\delta \mathbf {A} (\mathbf {s} )])>\ -<\ {\boldsymbol {i}}_{fc}(\mathbf {s} )|\ \delta \mathbf {A} (\mathbf {s} )>} (3-5-5)

式(3-5-5)から、二次の微小項を無視すると、Fの第”一”変分

δ F [ A ] = s R 3 <   ν ( s ) ( rot s [ A ( s ) ] )   |   ( rot s [ δ A ( s ) ] ) >   <   i f c ( s ) |   δ A ( s ) >   d 3 s {\displaystyle \delta F[A]={\int }_{\mathbf {s} \in \mathbb {R} ^{3}}<\ {\nu }(\mathbf {s} )(\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )])\ |\ (\operatorname {rot} _{\mathbf {s} }[\delta \mathbf {A} (\mathbf {s} )])>\ -<\ {\boldsymbol {i}}_{fc}(\mathbf {s} )|\ \delta \mathbf {A} (\mathbf {s} )>\ {d}^{3}\mathbf {s} } (3-5-6)

を得る。 式(3-5-6) に、ベクトル解析の公式を適用する。

div [ X × Y ] =<   Y   |   rot [ X ]   > <   X   |   rot [ Y ]   > {\displaystyle \operatorname {div} [\mathbf {X} \times \mathbf {Y} ]=<\ \mathbf {Y} \ |\ \operatorname {rot} [\mathbf {X} ]\ >-<\ \mathbf {X} \ |\ \operatorname {rot} [\mathbf {Y} ]\ >} (3-5-7)

において

X := ν ( s ) rot [ A ( s ) ] (3-5-8) Y := δ A (3-5-9) {\displaystyle {\begin{aligned}\mathbf {X} &:={\nu }(\mathbf {s} )\operatorname {rot} [\mathbf {A} (\mathbf {s} )]&{\text{(3-5-8)}}\\\mathbf {Y} &:=\delta \mathbf {A} &{\text{(3-5-9)}}\end{aligned}}}

を代入すると、

δ F [ A ] = s R 3 div s [ ( ν ( s ) rot s [ A ] ) × δ A ]   d 3 s + s R 3 < ( rot s [ ( ν ( s ) ) rot s [ A ( s ) ]   ] i f c ( s )   |   δ A ( s ) >   d 3 s (3-5-10) {\displaystyle {\begin{aligned}\delta F[\mathbf {A} ]&={\int }_{\mathbf {s} \in \mathbb {R} ^{3}}-\operatorname {div} _{\mathbf {s} }[({\nu }(\mathbf {s} )\operatorname {rot} _{\mathbf {s} }[\mathbf {A} ])\times \delta \mathbf {A} ]\ {d}^{3}\mathbf {s} \\&+{\int }_{\mathbf {s} \in \mathbb {R} ^{3}}<(\operatorname {rot} _{\mathbf {s} }[({\nu }(\mathbf {s} ))\operatorname {rot} _{\mathbf {s} }[\mathbf {A} (\mathbf {s} )]\ ]-{\boldsymbol {i}}_{fc}(\mathbf {s} )\ |\ \delta \mathbf {A} (\mathbf {s} )>\ {d}^{3}\mathbf {s} \qquad {\text{(3-5-10)}}\end{aligned}}}

式(3-5-10)の第一項については、式(3-4-6)と、「真空中ではdiv[H]=0であること」を考え併せると、 結局物質Ω内の効果しか寄与しないことが判る。さらにガウスの発散定理を考慮すると、式(3-5-10)の第一項は、

s Ω div s [ ( ν ( s ) rot s [ A ] ) × δ A   ]   d 3 s = s Ω ( ν ( s ) rot s [ A ] ) × δ A   d 2 ( Ω ) {\displaystyle {\int }_{\mathbf {s} \in \Omega }-\operatorname {div} _{\mathbf {s} }[({\nu }(\mathbf {s} )\operatorname {rot} _{\mathbf {s} }[\mathbf {A} ])\times \delta \mathbf {A} \ ]\ {d}^{3}\mathbf {s} =-{\int }_{\mathbf {s} \in \partial \Omega }({\nu }(\mathbf {s} )\operatorname {rot} _{\mathbf {s} }[\mathbf {A} ])\times \delta \mathbf {A} \ {d}^{2}(\partial \Omega )} (3-5-11)

となる。従って、式(3-5-10)の第一項が、任意の摂動δAに対して0になるためには、ノイマン条件即ち、

( rot [ A ] ) × n Ω   =   0 {\displaystyle (\operatorname {rot} [\mathbf {A} ])\times {n}_{\partial \Omega }\ =\ 0} (3-5-12)

が満たされればよい。さらに、式(3-5-10)の第二項が、任意の摂動δAに対して0になるためには、

rot [ ν rot [ A ] ] i f c = 0 {\displaystyle \operatorname {rot} [{\nu }\operatorname {rot} [\mathbf {A} ]]-{\boldsymbol {i}}_{fc}=0} (3-5-12)

であればよい。これは、「ベクトルポテンシャルによる静磁場の方程式」(静磁場の支配方程式)に(式(3-3-10))他ならない。以上から、ノイマン条件下での、静磁場の支配方程式の解は、汎関数Fの停留関数となること判る。

ベクトル解析に関する補足

静磁場の解析には、静磁場の解析には、やや高度なベクトル解析の知識 (例えば 藤本[33]参照) が要求される。同じ式変形が何度も繰り返されるのを避ける目的から、 一般的な知識は既知とする立場から、本記事の内容理解に必要と思われる事柄に限り、簡単に説明する。 が要求される。同じ式変形が何度も繰り返されるのを避ける目的から、 一般的な知識は既知とする立場から、本記事の内容理解に必要と思われる事柄に限り、簡単に説明する。

ベクトル場の代数演算と微分作用素

ベクトル場に代数演算を施したものに、微分作用素を作用させた場合に成り立つ公式について、本記事で用い、かつ、あまり本に載っていないものについて、簡潔にまとめる。

(1) <F|∇>について

F=(f1,f2,f3)をベクトル場とする。このとき、

  F   |     := f 1 x 1 + f 2 x 2 + f 3 x 3 {\displaystyle \left\langle \ \mathbf {F} \ |\ \nabla \ \right\rangle :={f}_{1}{\partial \over \partial {x}_{1}}+{f}_{2}{\partial \over \partial {x}_{2}}+{f}_{3}{\partial \over \partial {x}_{3}}}

と定義する。ここで、∇は、

:= x ^ x + y ^ y + z ^ z {\displaystyle \nabla :=\mathbf {\hat {x}} {\partial \over \partial x}+\mathbf {\hat {y}} {\partial \over \partial y}+\mathbf {\hat {z}} {\partial \over \partial z}}

を意味する。<F|∇>のことを、”F・∇”と書くこともある。

Gを、ベクトル場としたとき、<F|∇>をGに作用させると、

  F   |     G = f 1 g 1 x 1 + f 2 g 2 x 2 + f 3 g 3 x 3 = ( J [ G ] ) F {\displaystyle \left\langle \ \mathbf {F} \ |\ \nabla \ \right\rangle \mathbf {G} ={f}_{1}{\partial {g}_{1} \over \partial {x}_{1}}+{f}_{2}{\partial {g}_{2} \over \partial {x}_{2}}+{f}_{3}{\partial {g}_{3} \over \partial {x}_{3}}=(J[\mathbf {G} ])\cdot \mathbf {F} }

が成り立つ。ここで、 J [ G ] {\displaystyle J[\mathbf {G} ]} は、Gのヤコビ行列(ヤコビアンではない)を意味する。

(2)rot とスカラー倍、ベクトル積

F,Gを、ベクトル場、fを、スカラー値関数とする。このとき、以下が成り立つ。

rot [ f F ] = grad [ f ] × F + f rot [ F ] {\displaystyle \operatorname {rot} [f\mathbf {F} ]=\operatorname {grad} [f]\times \mathbf {F} +f\cdot \operatorname {rot} [\mathbf {F} ]}
rot [ F × G ] =   G   |     F   F   |     G + F div [ G ] G div [ F ] {\displaystyle \operatorname {rot} [\mathbf {F} \times \mathbf {G} ]=\left\langle \ \mathbf {G} \ |\ \nabla \ \right\rangle \mathbf {F} -\left\langle \ \mathbf {F} \ |\ \nabla \ \right\rangle \mathbf {G} +\mathbf {F} \operatorname {div} [\mathbf {G} ]-\mathbf {G} \operatorname {div} [\mathbf {F} ]}

(3)div とスカラー倍、ベクトル積

F,Gを、ベクトル場、fを、スカラー値関数とする。このとき、

div [ f F ] = F | grad [ f ] + f div [ F ] {\displaystyle \operatorname {div} \left[f\mathbf {F} \right]=\left\langle \mathbf {F} |\operatorname {grad} [f]\right\rangle +f\operatorname {div} [\mathbf {F} ]}

ディラックのデルタと、重積分

一般のn変数関数(関数はベクトル値関数であってよい)fの定義域をΩとしたとき、

f ( r ) = s Ω f ( s ) δ n ( r s )   d n s   = f δ n {\displaystyle f(\mathbf {r} )={\int }_{\mathbf {s} \in \Omega }f(\mathbf {s} ){\delta }^{n}(\mathbf {r} -\mathbf {s} )\ {d}^{n}s\ =f*{\delta }^{n}}

であることが知られている。ここで、上式の"*"は、合成積(乗算ではない)である。また、δnは、n変数のδ関数である。

発散微分とディラックのデルタ

原点(r=0)を除いて

div r [ r | r | ] = 3 | r | 3 3 | r | 2 | r | 5 = 0 {\displaystyle \operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} }{|\mathbf {r} |}}\right]={\frac {3}{|\mathbf {r} |^{3}}}-{\frac {3{|\mathbf {r} |}^{2}}{{|\mathbf {r} |}^{5}}}=0}
であり、原点(r=0)を中心とする、球体BLに対し、ガウスの発散定理を用いると、
r B L div r [ r | r | ]   d 3 r = 4 π {\displaystyle {\int }_{\mathbf {r} \in BL}\operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} }{|\mathbf {r} |}}\right]\ d^{3}r=4\pi }
となるので、結局、
div r [ r | r | ] = 4 π δ ( r ) {\displaystyle \operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} }{|\mathbf {r} |}}\right]=4\pi \delta (\mathbf {r} )}

であることが判る。(ここで、δは、3変数のδ関数である。
以上の議論を平行移動させると、

div r [ r s | r s | ] = 4 π δ ( r s ) {\displaystyle \operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |}}\right]=4\pi \delta (\mathbf {r} -\mathbf {s} )}

であることが判る。

分数関数の偏微分

s i [ 1 | r s | ] = s i [ 1 ( ( r i s i ) 2 + const. ) 1 / 2 ] = s i [ ( ( r i s i ) 2 + const. ) ( 1 / 2 ) ] {\displaystyle {\begin{aligned}{\frac {\partial }{\partial \mathbf {s} _{i}}}\left[{\frac {1}{|\mathbf {r} -\mathbf {s} |}}\right]&={\frac {\partial }{\partial {s}_{i}}}\left[{\frac {1}{{(({r}_{i}-{s}_{i})^{2}+{\text{const.}})}^{1/2}}}\right]\\&={\frac {\partial }{\partial {s}_{i}}}\left[{(({r}_{i}-{s}_{i})^{2}+{\text{const.}})}^{-(1/2)}\right]\end{aligned}}}

面積分に関する補足

ここでは、様々なベクトル場の面積分についてまとめる。 「絶対値による面積分」「外積面積分」という用語は一般的な用語ではないが、他に 適切な表現がないため、この場限りでそのような言い方をする。 本記事内での定義は、それぞれ以下の通り。

Iを R 2 {\displaystyle \mathbb {R} ^{2}} の閉長方形、
φ : I R 3 {\displaystyle \varphi :I\to \mathbb {R} ^{3}} は、 Iの近傍で、殆ど至る所区分的に滑らかかつ、非退化であり、かつ、Iの内部で単射なベクトル値関数、
S := φ [ I ] {\displaystyle S:=\varphi [I]} を、 R 3 {\displaystyle \mathbb {R} ^{3}} 内の曲面片とする。
また、Xを、Sの近傍で定義された、区分的に滑らかなベクトル場とする。このとき、

(1)普通の面積分

定義は、以下の通りである。
s S X ( s )   d 2 S := ( u 1 , u 2 ) I X ( φ ( u 1 , u 2 ) )   |   ( φ u 1 × φ u 2 )   d u 1 d u 2 {\displaystyle {\int }_{\mathbf {s} \in S}\mathbf {X} (\mathbf {s} )\ {d}^{2}S:={\int }_{({u}_{1},{u}_{2})\in I}\langle \mathbf {X} (\varphi ({u}_{1},{u}_{2}))\ |\ \left({\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right)\rangle \ d{u}_{1}d{u}_{2}}
上式の右辺は、 ( u 1 , u 2 ) {\displaystyle ({u}_{1},{u}_{2})} についてのスカラー値関数
X ( φ ( u 1 , u 2 ) )   |   ( φ u 1 × φ u 2 ) {\displaystyle \langle \mathbf {X} (\varphi ({u}_{1},{u}_{2}))\ |\ \left({\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right)\rangle }
を、区間I上で重積分したものを意味する。今、Sの単位法線ベクトル n S {\displaystyle {\mathbf {n} }_{S}} を、
n S := 1 | φ u 1 × φ u 2 | ( φ u 1 × φ u 2 ) {\displaystyle {\mathbf {n} }_{S}:={\frac {1}{\left|{\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right|}}\left({\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right)}
と定めると、
s S X ( s )   d 2 S = ( u 1 , u 2 ) I X ( φ ( u 1 , u 2 ) )   |   n S | φ u 1 × φ u 2 |   d u 1 d u 2 {\displaystyle {\int }_{\mathbf {s} \in S}\mathbf {X} (\mathbf {s} )\ {d}^{2}S={\int }_{({u}_{1},{u}_{2})\in I}\langle \mathbf {X} (\varphi ({u}_{1},{u}_{2}))\ |\ {\mathbf {n} }_{S}\rangle \left|{\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right|\ d{u}_{1}d{u}_{2}}
である。

(2)絶対値による面積分

定義は、以下の通りである。
s S X ( s )   | d 2 S | := ( u 1 , u 2 ) I X ( φ ( u 1 , u 2 ) )   | φ u 1 × φ u 2 |   d u 1 d u 2 {\displaystyle {\int }_{\mathbf {s} \in S}\mathbf {X} (\mathbf {s} )\ |{d}^{2}S|:={\int }_{({u}_{1},{u}_{2})\in I}\mathbf {X} (\varphi ({u}_{1},{u}_{2}))\ \left|{\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right|\ d{u}_{1}d{u}_{2}}
本記事では、絶対値による面積分の場合は、   | d 2 S | {\displaystyle \ |{d}^{2}S|} のように、面素に絶対値記号をつけることにする。
右辺は、 ( u 1 , u 2 ) {\displaystyle ({u}_{1},{u}_{2})} についてのベクトル値関数
X ( φ ( u 1 , u 2 ) )   | φ u 1 × φ u 2 | {\displaystyle \mathbf {X} (\varphi ({u}_{1},{u}_{2}))\ \left|{\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right|}
を、成分ごとに区間I上で重積分したものを意味する。即ち、絶対値による面積分は「関数の面積分を各成分ごとにやる」というのと同じ意味である。

(3)外積面積分

定義は、以下の通りである。
s S X ( s ) × d 2 S := ( u 1 , u 2 ) I ( X ( φ ( u 1 , u 2 ) ) ) × ( φ u 1 × φ u 2 )   d u 1 d u 2 {\displaystyle {\int }_{\mathbf {s} \in S}\mathbf {X} (\mathbf {s} )\times {d}^{2}S:={\int }_{({u}_{1},{u}_{2})\in I}\left(\mathbf {X} (\varphi ({u}_{1},{u}_{2}))\right)\times \left({\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{2}}}\right)\ d{u}_{1}d{u}_{2}}
本記事では、外積面積分の場合は、 ×   d 2 S {\displaystyle \times \ {d}^{2}S} のように、面素にの前に、×をつけることにする。
右辺は、 ( u 1 , u 2 ) {\displaystyle ({u}_{1},{u}_{2})} についてのベクトル値関数
X ( φ ( u 1 , u 2 ) ) × ( φ u 1 × φ u 1 ) {\displaystyle \mathbf {X} (\varphi ({u}_{1},{u}_{2}))\times \left({\frac {\partial \varphi }{\partial {u}_{1}}}\times {\frac {\partial \varphi }{\partial {u}_{1}}}\right)}
を、成分ごとに区間I上で重積分したものを意味する。即ち、(2)との関係でいえば、
s S X ( s ) × d 2 S = s S X ( s ) × n S | d 2 S | {\displaystyle {\int }_{\mathbf {s} \in S}\mathbf {X} (\mathbf {s} )\times {d}^{2}S={\int }_{\mathbf {s} \in S}\mathbf {X} (\mathbf {s} )\times \mathbf {n} _{S}|{d}^{2}S|}

となる。

脚注

[脚注の使い方]

注釈

  1. ^ 尚、磁化が Ω {\displaystyle \Omega } の外では0であるからといって、 Ω {\displaystyle \Omega } の外で A M {\displaystyle \mathbf {A} _{M}} が0となるとは限らない(大概の場合は Ω {\displaystyle \Omega } の外でも A M {\displaystyle \mathbf {A} _{M}} は、0ではない)ことに注意されたい。
  2. ^ なお、磁化は、 Ω {\displaystyle \Omega } の外では0なので、上式は、
    A M ( r ) = μ 0 4 π s R 3 ( ( M ( s ) ) × ( r s ) | r s | )   d 3 s {\displaystyle \mathbf {A} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \mathbb {R} ^{3}}\left({\frac {(\mathbf {M} (\mathbf {s} ))\times (\mathbf {r} -\mathbf {s} )}{|\mathbf {r} -\mathbf {s} |}}\right)\ {d}^{3}\mathbf {s} }
    としてもよい(紛らわしいのであまり推奨しない)。
  3. ^ 尚、磁化が Ω {\displaystyle \Omega } の外では0であるからといって、 Ω {\displaystyle \Omega } の外で B M {\displaystyle \mathbf {B} _{M}} が0となるとは限らない(大概の場合は Ω {\displaystyle \Omega } の外でも B M {\displaystyle \mathbf {B} _{M}} は、0ではない)ことに注意されたい。
  4. ^ なお、磁化は、 Ω {\displaystyle \Omega } の外では0なので、上式は、
    B M ( r ) = μ 0 4 π grad r [ M ( s ) | ( r s ) | r s | 3 ] + div r [ r s | r s | ] M ( s )   d 3 s {\displaystyle \mathbf {B} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]+\operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |}}\right]\mathbf {M} (\mathbf {s} )\ {d}^{3}\mathbf {s} }
    としてもよい(紛らわしいのであまり推奨しない)。
  5. ^ なお、磁化は、 Ω {\displaystyle \Omega } の外では0なので、上式は、
    M ( s )   =   s R 3   div r [ r s | r s | ] M ( s )   d 3 s {\displaystyle \mathbf {M} (\mathbf {s} )\ =\ {\int }_{\mathbf {s} \in \mathbf {R} ^{3}}\ \operatorname {div} _{\mathbf {r} }\left[{\frac {\mathbf {r} -\mathbf {s} }{|\mathbf {r} -\mathbf {s} |}}\right]\mathbf {M} (\mathbf {s} )\ {d}^{3}\mathbf {s} }
    でもある。
  6. ^ なお、磁化は、 Ω {\displaystyle \Omega } の外では0なので、上式は、
    B M ( r ) = μ 0 4 π s R 3 grad r [ M ( s ) | ( r s ) | r s | 3 ]   d 3 s + μ 0 M ( r ) {\displaystyle \mathbf {B} _{M}(\mathbf {r} )={\frac {\mu _{0}}{4\pi }}{\int }_{\mathbf {s} \in \mathbf {R} ^{3}}\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]\ {d}^{3}{s}+\mu _{0}\mathbf {M} (\mathbf {r} )}
    としてもよい(紛らわしいのであまり推奨しない)。
  7. ^ なお、磁化は、 Ω {\displaystyle \Omega } の外では0なので、上式は、
    H M ( r ) = 1 4 π s Ω grad r [ M ( s ) | ( r s ) | r s | 3 ]   d 3 s {\displaystyle \mathbf {H} _{M}(\mathbf {r} )={\frac {1}{4\pi }}{\int }_{\mathbf {s} \in \Omega }\operatorname {grad} _{\mathbf {r} }\left[{\frac {\langle \mathbf {M} (\mathbf {s} )|(\mathbf {r} -\mathbf {s} )\rangle }{{|\mathbf {r} -\mathbf {s} |}^{3}}}\right]\ {d}^{3}{s}}
    としてもよい(紛らわしいのであまり推奨しない)。
  8. ^ a b ここでいうδは、摂動あるいは変分を表し、ディラックのδと紛らわしいが全然別物である。

参考文献

  1. ^ W.K.H. パノフスキー (著), M. フィリップス (著),林 忠四郎 (翻訳), 西田 稔 (翻訳);「新版 電磁気学〈上〉」吉岡書店; POD版 (2002/09)
  2. ^ a b c d [★溝口 正 (著) 「電磁気学―SI UNITS 」裳華房 (2001/03)](特にP188付近)
  3. ^ [★]竹山 説三 (著) 「電磁気学現象理論」丸善出版; 3版 (1949)
  4. ^ a b [★]P.P.シルベスタ(著), R.L.フェラーリ(著), 本間 利久(著), 田中 康博(著); 「有限要素法による電磁界解析 (Information & computing (26)) 」 サイエンス社 (1988/09)
    [原書]Peter P. Silvester (Author), Ronald L. Ferrari (Author); 「Finite Elements for Electrical Engineers」Cambridge University Press; 3版 (1996/9/5) [1]
  5. ^ 斎藤兆古,坂本禎智,藤原耕二;「磁気回路法と有限要素法の理論的関係」 (PDF) 電気学会マグネティックス研究会資料 MAG-03 号:54-64 ページ:5-10 発行年:2003年03月31日
  6. ^ 静磁場解析のための二次要素を用いる有限要素法の研究
  7. ^ 有限要素法による磁場解析
  8. ^ a b 磁場が満たす偏微分方程式(秋田高専講義録)
  9. ^ 園田英徳;「大学院生のための基礎物理学」講談社(2011/09/29)ISBN 978-4-06-153277-9
  10. ^ 平川浩正;「電磁気学(新物理学シリーズ2)」培風館 (1986/04) ISBN 9784563024024
  11. ^ 守末 利弥「数値電磁気学のためのゲージ理論」森北出版 (1996/04) ISBN 978-4-627-71600-1
  12. ^ 依田 潔 (著) ;「Mathematicaによる電磁界シミュレーション入門 - POD版 (計算電気・電子工学シリーズ)」森北出版(2012/2/24) ISBN 978-4-627-71529-5
  13. ^ 電磁気学II(大阪大学 田中実教授の講義録) (PDF)
  14. ^ [2]立教大学講義ノート 本講義の参考文献欄 等、至る所に文字化けがあるが、たとえば、もじばけらった等によって解読できる。
  15. ^ 東京理科大学講義ノート (PDF) および、講義スライド (PDF)
  16. ^ 鹿児島 誠一 (著) ;「電磁気学 (パリティ物理学コース)」丸善(1997/01) ISBN 978-4-621-04277-9
  17. ^ 後藤 憲一(著), 山崎修一郎(著)  ;「詳解 電磁気学演習」共立出版 (1970/12) ISBN 978-4-320-03022-0
  18. ^ 前田 和茂 (著) ,小林 俊雄(著);「ビジュアルアプローチ電磁気学」森北出版 (2009/12/5) ISBN 978-4-627-16221-1
  19. ^ Julius Adams Stratton;「Electromagnetic Theory」Wiley-IEEE Press; 1版 (2007/1/22) ISBN 978-0-470-13153-4
  20. ^ A. Pramanik (著) ;「Electro-Magnetism: Theory and Applications 」Prentice-Hall of India Pvt.Ltd (2004/8/15) [3]
  21. ^ Bo-nan Jiang(著) ;「The Least-Squares Finite Element Method: Theory and Applications in Computational Fluid Dynamics and Electromagnetics (Scientific Computation) 」Springer; 1998版 (1998/6/22) [4]
  22. ^ 電気磁気学特論(秋田高専講義録)
  23. ^ 高橋 康人 (早稲田大学):"高速多重極法を組み込んだ磁気モーメント法による 磁性体解析に関する基礎的検討" (PDF)
  24. ^ 東京電機大学 (編集)「入門 電磁気学」東京電機大学出版局 (2006/03) ISBN 9784501004200
  25. ^ 早川 義晴 (著) 「電気教科書 電験三種合格ガイド」翔泳社 (2011/2/25) ISBN 9784798126623
  26. ^ 磁気回路と電気回路 (PDF)
  27. ^ 坪井 一洋 (著) 「システムと微分方程式」三恵社 (2011/5/22) ISBN 9784883618248
  28. ^ 複数磁石による静磁場(簡易シミュレータ)
  29. ^ 近畿大学 講義ノート (PDF)
  30. ^ 等々力 二郎:「有限の太さの矩形断面ヘリカルコイルの磁場の計算」核融合研究,57(1987)318
  31. ^ 渡辺 二太:「空心円筒コイル群によって作られる磁場の計算」 核融合研究 Vol.35 (1976) No.3 P235-242
  32. ^ 渡辺 二太:「多様な形状のコイルに対する磁場計算法」 核融合研究 Vol.63 (1990) No.6 P482-507
  33. ^ 藤本 淳夫 (著) ベクトル解析 (現代数学レクチャーズ C- 1) 培風館 (1979/01)

関連項目

外部リンク

  • 磁化Mと磁場Hについて (PDF)
  • 磁場 (PDF)
  • 電気磁気学演習解答例 (PDF)