10.1

Problem 10.1

schwarzChapter 10

习题 10.1

来源: 第10章, PDF第181,182页


10.1 We saw that the Dirac equation predicted that there is an interaction between the electron spin and the magnetic field, SB\vec{S}\vec{B}, with strength μB=e2mec\mu_B = \frac{\hbar e}{2m_e c}. When the electron has angular momentum L\vec{L}, such as in an atomic orbital, there is also a BL\vec{B}\vec{L} interaction and a spin-orbit coupling SL\vec{S}\vec{L}. The Dirac equation (along with symmetry arguments) predicts the strength of all three interactions, as well as other corrections. To see the effect of these terms on the hydrogen atom, we have to take the non-relativistic limit.

(a) For the Schrödinger equation, we need the Hamiltonian, not the Lagrangian. Find the Dirac Hamiltonian by writing the Dirac equation as itψ=HDψi\partial_t \psi = H_D \psi. Write the Hamiltonian in terms of momenta pip_i rather than derivatives i\partial_i.

(b) Calculate (HDeA0)2(H_D - eA_0)^2 in the Weyl representation for ψ=(ψLψR)\psi = (\psi_L \psi_R). Leave in terms of σi\sigma_i, pip_i and AiA_i. Put back in the factors of cc and \hbar, keeping the charge ee dimensionless.

(c) Now take the square root of this result and expand in 1c\frac{1}{c}, subtracting off the zero-point energy mc2mc^2, i.e. compute H=HDmc2H = H_D - mc^2 to order c0c^0. Looking at the σi\sigma_i term, how big are the electron's electric and magnetic dipole moments?

(d) The size of the terms in this Hamiltonian are only meaningful because the spin and angular momentum operators have the same normalization. Check the normalization of the angular momentum operators Li=εijkxjpkL_i = \varepsilon_{ijk} x_j p_k and the spin operators Si=12σiS_i = \frac{1}{2} \sigma_i by showing that they both satisfy the rotation algebra: [Ji,Jj]=iεijkJk[J_i, J_j] = i\varepsilon_{ijk} J_k.

(e) The gyromagnetic ratio, geg_e (sometimes called the gg-factor), is the relative size of the SB\vec{S}\vec{B} and LB\vec{L}\vec{B} interactions. Choose a constant magnetic field in the zz direction, then isolate the BzLzB_z L_z coupling in HH. Extract the electron gyromagnetic ratio geg_e by writing the entire coupling to the magnetic field in the Hamiltonian as μBBz(Lz+geSz)=Bzμz\mu_B B_z (L_z + g_e S_z) = B_z \mu_z, with μμB(L+geS)\vec{\mu} \equiv \mu_B (\vec{L} + g_e \vec{S}). How could you experimentally measure geg_e (e.g. with spectroscopy of the hydrogen atom)?

(f) In spherical coordinates, the Schrödinger equation has an L2\vec{L}^2 term. With spin, you might expect that this becomes Lμ=μB(L2+geLS)\vec{L}\vec{\mu} = \mu_B (\vec{L}^2 + g_e \vec{L}\vec{S}), making the LS\vec{L}\vec{S} term proportional to the gg-factor. This is wrong. It misses an important relativistic effect, Thomas precession. It is very hard to calculate directly, but easy to calculate using symmetries. With no magnetic field, the atom, with spin included, is

still rotationally invariant. Which of J=L+S\vec{J} = \vec{L} + \vec{S} or μ=L+geS\vec{\mu} = \vec{L} + g_e \vec{S} is conserved (i.e. commutes with HH)? Using this result, how does the spin-orbit coupling depend on geg_e?

(g) There are additional relativistic effects coming from the Dirac equation. Expand the Dirac equation to next order in 1c2\frac{1}{c^2}, producing a term that scales as p4\vec{p}^4.

(h) Now let us do some dimensional analysis - there is only one scale mem_e. Show that the electron's Compton wavelength, the classical electron radius, rer_e, the Bohr radius, a0a_0, and the inverse-Rydberg constant, Ry1\text{Ry}^{-1}, are all me1m_e^{-1} times powers of αe\alpha_e. Are the splittings due to the p4p^4 term fine structure (ΔEαe2E\Delta E \sim \alpha_e^2 E), hyperfine structure (ΔEαe4E\Delta E \sim \alpha_e^4 E) or something else? [Hint: write out a formula for the energy shift using time-independent perturbation theory, then see which of the above length scales appears.]

习题 10.1 - 解答


(a) 狄拉克方程在电磁场中的协变形式为 (iγμDμm)ψ=0(i\gamma^\mu D_\mu - m)\psi = 0,其中协变导数 Dμ=μ+ieAμD_\mu = \partial_\mu + ieA_\mu。将其展开并分离出时间导数项: iγ0(0+ieA0)ψ+iγk(k+ieAk)ψmψ=0i\gamma^0 (\partial_0 + ieA_0)\psi + i\gamma^k (\partial_k + ieA_k)\psi - m\psi = 0 左乘 γ0\gamma^0 并利用 (γ0)2=1(\gamma^0)^2 = 1,得到: i(0+ieA0)ψ+iγ0γk(k+ieAk)ψγ0mψ=0i(\partial_0 + ieA_0)\psi + i\gamma^0\gamma^k (\partial_k + ieA_k)\psi - \gamma^0 m\psi = 0 将时间偏导数移到等式左边,并代入动量算符 pk=ikp_k = -i\partial_k(即 k=ipk\partial_k = ip_k): itψ=[eA0+γ0γk(pkeAk)+γ0m]ψi\partial_t \psi = \left[ eA_0 + \gamma^0\gamma^k (p_k - eA_k) + \gamma^0 m \right] \psi 定义狄拉克矩阵 αk=γ0γk\alpha^k = \gamma^0\gamma^kβ=γ0\beta = \gamma^0,即可得到狄拉克哈密顿量: HD=α(peA)+βm+eA0\boxed{H_D = \vec{\alpha} \cdot (\vec{p} - e\vec{A}) + \beta m + eA_0}

(b) 在 Weyl 表象中,狄拉克矩阵的形式为 γ0=(0II0)\gamma^0 = \begin{pmatrix} 0 & I \\ I & 0 \end{pmatrix}γ=(0σσ0)\vec{\gamma} = \begin{pmatrix} 0 & \vec{\sigma} \\ -\vec{\sigma} & 0 \end{pmatrix}。 因此 α=γ0γ=(σ00σ)\vec{\alpha} = \gamma^0\vec{\gamma} = \begin{pmatrix} \vec{\sigma} & 0 \\ 0 & -\vec{\sigma} \end{pmatrix}β=(0II0)\beta = \begin{pmatrix} 0 & I \\ I & 0 \end{pmatrix}。 令 π=peA\vec{\pi} = \vec{p} - e\vec{A},则 HDeA0H_D - eA_0 的矩阵形式为: HDeA0=(σπmmσπ)H_D - eA_0 = \begin{pmatrix} \vec{\sigma}\cdot\vec{\pi} & m \\ m & -\vec{\sigma}\cdot\vec{\pi} \end{pmatrix} 计算其平方: (HDeA0)2=((σπ)2+m200(σπ)2+m2)(H_D - eA_0)^2 = \begin{pmatrix} (\vec{\sigma}\cdot\vec{\pi})^2 + m^2 & 0 \\ 0 & (\vec{\sigma}\cdot\vec{\pi})^2 + m^2 \end{pmatrix} 利用泡利矩阵的恒等式 σiσj=δij+iϵijkσk\sigma_i \sigma_j = \delta_{ij} + i\epsilon_{ijk}\sigma_k,展开 (σπ)2(\vec{\sigma}\cdot\vec{\pi})^2(σπ)2=σiπiσjπj=π2+iϵijkσkπiπj(\vec{\sigma}\cdot\vec{\pi})^2 = \sigma_i \pi_i \sigma_j \pi_j = \vec{\pi}^2 + i\epsilon_{ijk}\sigma_k \pi_i \pi_j 由于 [πi,πj]=[pieAi,pjeAj]=e[pi,Aj]e[Ai,pj]=ie(iAjjAi)=ieϵijmBm[\pi_i, \pi_j] = [p_i - eA_i, p_j - eA_j] = -e[p_i, A_j] - e[A_i, p_j] = i\hbar e (\partial_i A_j - \partial_j A_i) = i\hbar e \epsilon_{ijm} B_m, 代入上式得到交叉项: iϵijkσkπiπj=i2ϵijkσk[πi,πj]=i2ϵijkσk(ieϵijmBm)=eσBi\epsilon_{ijk}\sigma_k \pi_i \pi_j = \frac{i}{2}\epsilon_{ijk}\sigma_k [\pi_i, \pi_j] = \frac{i}{2}\epsilon_{ijk}\sigma_k (i\hbar e \epsilon_{ijm} B_m) = -e\hbar \vec{\sigma}\cdot\vec{B} 恢复光速 cc(动量项乘 cc,质量项乘 c2c^2),得到: (HDeA0)2=c2(peA)2ec2σB+m2c4\boxed{(H_D - eA_0)^2 = c^2(\vec{p} - e\vec{A})^2 - e\hbar c^2 \vec{\sigma}\cdot\vec{B} + m^2 c^4}

(c) 对上式开平方并提取 mc2mc^2 因子: HDeA0=mc21+(peA)2m2c2eσBm2c2H_D - eA_0 = mc^2 \sqrt{1 + \frac{(\vec{p}-e\vec{A})^2}{m^2 c^2} - \frac{e\hbar \vec{\sigma}\cdot\vec{B}}{m^2 c^2}} 在非相对论极限下(1/c1/c 展开),利用 1+x1+x2\sqrt{1+x} \approx 1 + \frac{x}{2}HDeA0+mc2+(peA)22me2mσBH_D \approx eA_0 + mc^2 + \frac{(\vec{p}-e\vec{A})^2}{2m} - \frac{e\hbar}{2m}\vec{\sigma}\cdot\vec{B} 减去静止能量 mc2mc^2,得到 c0c^0 阶的哈密顿量: H=eA0+(peA)22me2mσB\boxed{H = eA_0 + \frac{(\vec{p}-e\vec{A})^2}{2m} - \frac{e\hbar}{2m}\vec{\sigma}\cdot\vec{B}} 哈密顿量中与电磁场耦合的偶极矩项形式为 dEEμMB-\vec{d}_E \cdot \vec{E} - \vec{\mu}_M \cdot \vec{B}。 观察 σi\sigma_i 项可知,电子的电偶极矩为 0\boxed{0}(无 σE\vec{\sigma}\cdot\vec{E} 项),磁偶极矩大小为 μM=e2m\boxed{\mu_M = \frac{e\hbar}{2m}}

(d) 为了验证代数关系,设 =1\hbar=1。对于轨道角动量 Li=ϵijkxjpkL_i = \epsilon_{ijk} x_j p_k[Li,Lj]=ϵiabϵjcd[xapb,xcpd][L_i, L_j] = \epsilon_{iab}\epsilon_{jcd} [x_a p_b, x_c p_d] 利用 [xa,pd]=iδad[x_a, p_d] = i\delta_{ad}[pb,xc]=iδbc[p_b, x_c] = -i\delta_{bc}[xapb,xcpd]=xa[pb,xc]pd+xc[xa,pd]pb=ixapdδbc+ixcpbδad[x_a p_b, x_c p_d] = x_a [p_b, x_c] p_d + x_c [x_a, p_d] p_b = -i x_a p_d \delta_{bc} + i x_c p_b \delta_{ad} 代回对易子并利用列维-奇维塔符号的缩并公式 ϵiabϵjcb=δijδacδicδaj\epsilon_{iab}\epsilon_{jcb} = \delta_{ij}\delta_{ac} - \delta_{ic}\delta_{aj}[Li,Lj]=i(δijδacδicδaj)xapc+i(δijδbdδidδbj)xcpb=i(xipjxjpi)[L_i, L_j] = -i(\delta_{ij}\delta_{ac} - \delta_{ic}\delta_{aj})x_a p_c + i(\delta_{ij}\delta_{bd} - \delta_{id}\delta_{bj})x_c p_b = i(x_i p_j - x_j p_i)iϵijkLk=iϵijkϵkabxapb=i(δiaδjbδibδja)xapb=i(xipjxjpi)i\epsilon_{ijk}L_k = i\epsilon_{ijk}\epsilon_{kab}x_a p_b = i(\delta_{ia}\delta_{jb} - \delta_{ib}\delta_{ja})x_a p_b = i(x_i p_j - x_j p_i)。 故 [Li,Lj]=iϵijkLk\boxed{[L_i, L_j] = i\epsilon_{ijk}L_k}

对于自旋算符 Si=12σiS_i = \frac{1}{2}\sigma_i[Si,Sj]=14[σi,σj]=14(2iϵijkσk)=iϵijk(12σk)[S_i, S_j] = \frac{1}{4}[\sigma_i, \sigma_j] = \frac{1}{4}(2i\epsilon_{ijk}\sigma_k) = i\epsilon_{ijk}\left(\frac{1}{2}\sigma_k\right)[Si,Sj]=iϵijkSk\boxed{[S_i, S_j] = i\epsilon_{ijk}S_k}。两者均满足标准旋转代数。

(e) 取恒定磁场 B=Bzz^\vec{B} = B_z \hat{z},选择对称规范 A=12B×r\vec{A} = \frac{1}{2}\vec{B} \times \vec{r}。 展开动能项中的交叉项(库仑规范下 A=0\nabla\cdot\vec{A}=0,故 pA=Ap\vec{p}\cdot\vec{A} = \vec{A}\cdot\vec{p}): e2m(pA+Ap)=emAp=e2m(B×r)p=e2mB(r×p)=e2mBL=e2mBzLz-\frac{e}{2m}(\vec{p}\cdot\vec{A} + \vec{A}\cdot\vec{p}) = -\frac{e}{m}\vec{A}\cdot\vec{p} = -\frac{e}{2m}(\vec{B}\times\vec{r})\cdot\vec{p} = -\frac{e}{2m}\vec{B}\cdot(\vec{r}\times\vec{p}) = -\frac{e}{2m}\vec{B}\cdot\vec{L} = -\frac{e}{2m}B_z L_z 自旋磁耦合项为 e2mσB=emSB=emBzSz-\frac{e}{2m}\vec{\sigma}\cdot\vec{B} = -\frac{e}{m}\vec{S}\cdot\vec{B} = -\frac{e}{m}B_z S_z。 总磁耦合哈密顿量为: Hmag=e2mBz(Lz+2Sz)H_{mag} = -\frac{e}{2m}B_z (L_z + 2S_z) 定义玻尔磁子 μB=e2m\mu_B = \frac{e}{2m}(取电荷大小),对比形式 μBBz(Lz+geSz)\mu_B B_z (L_z + g_e S_z),可提取出: ge=2\boxed{g_e = 2} 实验测量方法:可以通过氢原子的反常塞曼效应(Anomalous Zeeman effect)光谱学来测量。将氢原子置于外加磁场中,测量能级分裂的谱线间距,由于分裂正比于 μBBz(ml+gems)\mu_B B_z (m_l + g_e m_s),从中即可拟合提取出 geg_e 的值。

(f) 在没有外加磁场的情况下,孤立原子的空间旋转对称性未被破坏。 由于 J\vec{J} 是空间旋转的生成元,因此总角动量 J=L+S\boxed{\vec{J} = \vec{L} + \vec{S}} 是守恒的(与 HH 对易),而 μ\vec{\mu} 不守恒。 因为 J\vec{J} 守恒,自旋-轨道耦合项必须是旋转标量,即必须正比于 LS\vec{L}\cdot\vec{S}。由于磁矩相互作用给出的系数正比于 geg_e,而 Thomas 进动(纯相对论运动学效应)会产生一个方向相反、大小相当于 ge=1g_e=1 的修正项,因此自旋-轨道耦合的强度依赖于 (ge1)\boxed{(g_e - 1)}

(g) 为了得到 p4\vec{p}^4 项,我们在无电磁场(A=0,B=0\vec{A}=0, \vec{B}=0)的情况下将狄拉克哈密顿量展开到 1/c21/c^2 的下一阶。 利用泰勒展开 1+x1+12x18x2\sqrt{1+x} \approx 1 + \frac{1}{2}x - \frac{1}{8}x^2HD=mc21+p2m2c2mc2(1+p22m2c2p48m4c4)=mc2+p22mp48m3c2H_D = mc^2 \sqrt{1 + \frac{\vec{p}^2}{m^2 c^2}} \approx mc^2 \left( 1 + \frac{\vec{p}^2}{2m^2 c^2} - \frac{\vec{p}^4}{8m^4 c^4} \right) = mc^2 + \frac{\vec{p}^2}{2m} - \frac{\vec{p}^4}{8m^3 c^2} 因此,缩放为 p4\vec{p}^4 的相对论修正项为: Hp4=p48m3c2\boxed{H_{p^4} = -\frac{\vec{p}^4}{8m^3 c^2}}

(h) 在自然单位制(或仅关注量纲与 αe\alpha_e 的幂次)下,唯一的基本质量尺度是 mem_e

  1. 康普顿波长 λC=mecme1\lambda_C = \frac{\hbar}{m_e c} \propto m_e^{-1},故 λCαe0me1\boxed{\lambda_C \sim \alpha_e^0 m_e^{-1}}
  2. 经典电子半径 re=e24πmec2=αemecr_e = \frac{e^2}{4\pi m_e c^2} = \alpha_e \frac{\hbar}{m_e c},故 reαe1me1\boxed{r_e \sim \alpha_e^1 m_e^{-1}}
  3. 玻尔半径 a0=αemeca_0 = \frac{\hbar}{\alpha_e m_e c},故 a0αe1me1\boxed{a_0 \sim \alpha_e^{-1} m_e^{-1}}
  4. 逆里德伯常数 Ry1=(12mec2αe2)11meαe2\text{Ry}^{-1} = \left( \frac{1}{2} m_e c^2 \alpha_e^2 \right)^{-1} \propto \frac{1}{m_e \alpha_e^2},故 Ry1αe2me1\boxed{\text{Ry}^{-1} \sim \alpha_e^{-2} m_e^{-1}}

能级移动分析: 利用不含时微扰论,能量移动为 ΔE=p48me3c2\Delta E = \langle -\frac{p^4}{8m_e^3 c^2} \rangle。 在氢原子中,动量的特征尺度由玻尔半径决定:pa0αemecp \sim \frac{\hbar}{a_0} \sim \alpha_e m_e c。 代入微扰项: ΔE(αemec)4me3c2=αe4mec2\Delta E \sim \frac{(\alpha_e m_e c)^4}{m_e^3 c^2} = \alpha_e^4 m_e c^2 而未微扰的基态能量(里德伯能量)尺度为 ERyαe2mec2E \sim \text{Ry} \sim \alpha_e^2 m_e c^2。 因此,能量移动与原能级的关系为 ΔEαe2(αe2mec2)αe2E\Delta E \sim \alpha_e^2 (\alpha_e^2 m_e c^2) \sim \alpha_e^2 E。 这表明 p4p^4 项引起的能级分裂属于 精细结构 (Fine structure)\boxed{\text{精细结构 (Fine structure)}}

10.2

Problem 10.2

schwarzChapter 10

习题 10.2

来源: 第10章, PDF第182页


10.2 In this problem you will construct the finite-dimensional irreducible representations of SU(2)\text{SU}(2). By definition, such a representation is a set of three n×nn \times n matrices τ1,τ2\tau_1, \tau_2 and τ3\tau_3 satisfying the algebra of the Pauli matrices [τi,τj]=iεijkτk[\tau_i, \tau_j] = i\varepsilon_{ijk} \tau_k. It is also helpful to define the linear combinations τ±=τ1±iτ2\tau^{\pm} = \tau_1 \pm i\tau_2.

(a) In any such representation we can diagonalize τ3\tau_3. Its eigenvectors are nn complex vectors VjV_j with τ3Vj=λjVj\tau_3 V_j = \lambda_j V_j. Show that τ+Vj\tau^+ V_j and τVj\tau^- V_j either vanish or are eigenstates of τ3\tau_3 with eigenvalues λj+1\lambda_j + 1 and λj1\lambda_j - 1 respectively.

(b) Prove that exactly one of the eigenstates VmaxV_{\max} of τ3\tau_3 must satisfy τ+Vmax=0\tau^+ V_{\max} = 0. The eigenvalue λmax=j\lambda_{\max} = j of VmaxV_{\max} is known as the spin. Similarly, there will be an eigenvector VminV_{\min} of τ3\tau_3 with τVmin=0\tau^- V_{\min} = 0.

(c) Since there are a finite number of eigenvectors, Vmin=(τ)NVmaxV_{\min} = (\tau^-)^N V_{\max} for some integer NN. Prove that N=2JN = 2J so that n=2J+1n = 2J + 1.

(d) Construct explicitly the five-dimensional representation of SU(2)\text{SU}(2).

习题 10.2 - 解答


习题 10.2 分析与解答

首先,由 SU(2)\text{SU}(2) 李代数生成元的对易关系 [τi,τj]=iεijkτk[\tau_i, \tau_j] = i\varepsilon_{ijk} \tau_k,我们可以推导出升降算符 τ±=τ1±iτ2\tau^{\pm} = \tau_1 \pm i\tau_2τ3\tau_3 的对易关系: [τ3,τ±]=[τ3,τ1]±i[τ3,τ2]=iτ2±i(iτ1)=±(τ1±iτ2)=±τ±[\tau_3, \tau^{\pm}] = [\tau_3, \tau_1] \pm i[\tau_3, \tau_2] = i\tau_2 \pm i(-i\tau_1) = \pm(\tau_1 \pm i\tau_2) = \pm \tau^{\pm} [τ+,τ]=[τ1+iτ2,τ1iτ2]=i[τ1,τ2]+i[τ2,τ1]=2τ3[\tau^+, \tau^-] = [\tau_1 + i\tau_2, \tau_1 - i\tau_2] = -i[\tau_1, \tau_2] + i[\tau_2, \tau_1] = 2\tau_3 这些基本对易关系是后续推导的核心。


(a)

已知 τ3Vj=λjVj\tau_3 V_j = \lambda_j V_j。我们需要考察 τ3\tau_3 作用在态 τ±Vj\tau^{\pm} V_j 上的结果。利用对易关系 [τ3,τ±]=±τ±[\tau_3, \tau^{\pm}] = \pm \tau^{\pm},即 τ3τ±=τ±τ3±τ±\tau_3 \tau^{\pm} = \tau^{\pm} \tau_3 \pm \tau^{\pm},可得: τ3(τ±Vj)=(τ±τ3±τ±)Vj=τ±(λjVj)±τ±Vj=(λj±1)(τ±Vj)\tau_3 (\tau^{\pm} V_j) = (\tau^{\pm} \tau_3 \pm \tau^{\pm}) V_j = \tau^{\pm} (\lambda_j V_j) \pm \tau^{\pm} V_j = (\lambda_j \pm 1) (\tau^{\pm} V_j) 这表明,如果 τ±Vj0\tau^{\pm} V_j \neq 0,那么它必然是 τ3\tau_3 的本征态,且对应的本征值为 λj±1\lambda_j \pm 1τ3(τ±Vj)=(λj±1)(τ±Vj)\boxed{\tau_3 (\tau^{\pm} V_j) = (\lambda_j \pm 1) (\tau^{\pm} V_j)}


(b)

由于该表示是有限维的(n×nn \times n 矩阵),τ3\tau_3 只能有有限个本征值。 如果对于所有的本征态 VjV_j 都有 τ+Vj0\tau^+ V_j \neq 0,那么通过不断作用 τ+\tau^+,我们将得到一个具有无限个本征值 λj,λj+1,λj+2,\lambda_j, \lambda_j+1, \lambda_j+2, \dots 的本征态序列,这与表示的有限维性矛盾。 因此,必然存在至少一个具有最大本征值 λmaxj\lambda_{\max} \equiv j 的态 VmaxV_{\max},使得升算符作用其上得到零向量: τ+Vmax=0\tau^+ V_{\max} = 0 进一步,由于我们讨论的是不可约表示(irreducible representation),整个表示空间必须能由单一的最高权态通过降算符 τ\tau^- 的反复作用生成。如果存在两个线性无关的最高权态,它们将各自生成一个不变子空间,这与不可约性矛盾。因此,满足 τ+Vmax=0\tau^+ V_{\max} = 0 的最高权态 VmaxV_{\max} 是唯一的(不考虑常数因子)。 同理,必然存在唯一的最低权态 VminV_{\min},满足 τVmin=0\tau^- V_{\min} = 0


(c)

设最高权态为 V0VmaxV_0 \equiv V_{\max},其本征值为 jj。定义一系列态 Vk=(τ)kV0V_k = (\tau^-)^k V_0。 我们需要计算 τ+\tau^+ 作用在 VkV_k 上的结果。假设 τ+Vk=ckVk1\tau^+ V_k = c_k V_{k-1},我们通过数学归纳法求解系数 ckc_k。 当 k=0k=0 时,τ+V0=0\tau^+ V_0 = 0,故 c0=0c_0 = 0。 对于 k>0k > 0τ+Vk=τ+τVk1=([τ+,τ]+ττ+)Vk1=(2τ3+ττ+)Vk1\tau^+ V_k = \tau^+ \tau^- V_{k-1} = ([\tau^+, \tau^-] + \tau^- \tau^+) V_{k-1} = (2\tau_3 + \tau^- \tau^+) V_{k-1} 由于 Vk1V_{k-1}τ3\tau_3 的本征态,本征值为 j(k1)j - (k-1),代入上式得: ckVk1=2(jk+1)Vk1+τ(ck1Vk2)=[2(jk+1)+ck1]Vk1c_k V_{k-1} = 2(j - k + 1) V_{k-1} + \tau^- (c_{k-1} V_{k-2}) = [2(j - k + 1) + c_{k-1}] V_{k-1} 由此得到递推关系:ckck1=2(jk+1)c_k - c_{k-1} = 2(j - k + 1)。 求和可得: ck=m=1k2(jm+1)=2kj2k(k+1)2+2k=k(2jk+1)c_k = \sum_{m=1}^k 2(j - m + 1) = 2kj - 2\frac{k(k+1)}{2} + 2k = k(2j - k + 1) 由于表示是有限维的,降算符序列必然在某一步终止。设 VNV_N 是最后一个非零态(即 VminV_{\min}),则 VN+1=τVN=0V_{N+1} = \tau^- V_N = 0。 既然 VN+1=0V_{N+1} = 0,必然有 τ+VN+1=0\tau^+ V_{N+1} = 0,即 cN+1=0c_{N+1} = 0。代入 ckc_k 的表达式: cN+1=(N+1)(2j(N+1)+1)=(N+1)(2jN)=0c_{N+1} = (N+1)(2j - (N+1) + 1) = (N+1)(2j - N) = 0 因为 N0N \ge 0,所以 N+10N+1 \neq 0,必然有 2jN=02j - N = 0,即: N=2j\boxed{N = 2j} 态的序列为 V0,V1,,V2jV_0, V_1, \dots, V_{2j},总共有 2j+12j + 1 个线性无关的态,因此表示的维数为 n=2j+1\boxed{n = 2j + 1}


(d)

对于 5 维表示,n=5n = 5,因此 2j+1=5    j=22j + 1 = 5 \implies j = 2。 基矢可以标记为 j,m|j, m\rangle,其中 m=2,1,0,1,2m = 2, 1, 0, -1, -2。 在这一组基下,τ3\tau_3 是对角矩阵: τ3=(2000001000000000001000002)\tau_3 = \begin{pmatrix} 2 & 0 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & -1 & 0 \\ 0 & 0 & 0 & 0 & -2 \end{pmatrix} 升降算符的矩阵元由标准公式给出(注意归一化系数 ck\sqrt{c_k}): τ±j,m=j(j+1)m(m±1)j,m±1\tau^{\pm} |j, m\rangle = \sqrt{j(j+1) - m(m \pm 1)} |j, m \pm 1\rangle 对于 j=2j=2,计算非零矩阵元:

  • τ+2,1=622,2=22,2\tau^+ |2, 1\rangle = \sqrt{6-2} |2, 2\rangle = 2 |2, 2\rangle
  • τ+2,0=602,1=62,1\tau^+ |2, 0\rangle = \sqrt{6-0} |2, 1\rangle = \sqrt{6} |2, 1\rangle
  • τ+2,1=602,0=62,0\tau^+ |2, -1\rangle = \sqrt{6-0} |2, 0\rangle = \sqrt{6} |2, 0\rangle
  • τ+2,2=622,1=22,1\tau^+ |2, -2\rangle = \sqrt{6-2} |2, -1\rangle = 2 |2, -1\rangle

由此写出 τ+\tau^+ 矩阵,而 τ\tau^-τ+\tau^+ 的转置(共轭转置)。进而利用 τ1=12(τ++τ)\tau_1 = \frac{1}{2}(\tau^+ + \tau^-)τ2=12i(τ+τ)\tau_2 = \frac{1}{2i}(\tau^+ - \tau^-),可显式构造出 τ1\tau_1τ2\tau_2

τ1=(01000103200032032000320100010)\boxed{ \tau_1 = \begin{pmatrix} 0 & 1 & 0 & 0 & 0 \\ 1 & 0 & \sqrt{\frac{3}{2}} & 0 & 0 \\ 0 & \sqrt{\frac{3}{2}} & 0 & \sqrt{\frac{3}{2}} & 0 \\ 0 & 0 & \sqrt{\frac{3}{2}} & 0 & 1 \\ 0 & 0 & 0 & 1 & 0 \end{pmatrix} }

τ2=(0i000i0i32000i320i32000i320i000i0)\boxed{ \tau_2 = \begin{pmatrix} 0 & -i & 0 & 0 & 0 \\ i & 0 & -i\sqrt{\frac{3}{2}} & 0 & 0 \\ 0 & i\sqrt{\frac{3}{2}} & 0 & -i\sqrt{\frac{3}{2}} & 0 \\ 0 & 0 & i\sqrt{\frac{3}{2}} & 0 & -i \\ 0 & 0 & 0 & i & 0 \end{pmatrix} }

τ3=(2000001000000000001000002)\boxed{ \tau_3 = \begin{pmatrix} 2 & 0 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & -1 & 0 \\ 0 & 0 & 0 & 0 & -2 \end{pmatrix} } 这三个矩阵即为 SU(2)\text{SU}(2) 的 5 维不可约表示的显式构造。

10.3

Problem 10.3

schwarzChapter 10

习题 10.3

来源: 第10章, PDF第182页


10.3 Derive Eqs. (10.141) and (10.142):

(a) gμνσμαα˙σνββ˙=2εαβεα˙β˙g^{\mu\nu} \sigma_{\mu}^{\alpha\dot{\alpha}} \sigma_{\nu}^{\beta\dot{\beta}} = 2\varepsilon^{\alpha\beta} \varepsilon^{\dot{\alpha}\dot{\beta}},

(b) εαβεα˙β˙σμββ˙=σˉα˙αμ\varepsilon_{\alpha\beta} \varepsilon_{\dot{\alpha}\dot{\beta}} \sigma^{\mu\beta\dot{\beta}} = \bar{\sigma}_{\dot{\alpha}\alpha}^{\mu}.

习题 10.3 - 解答


本题要求推导 Weyl 旋量表示中泡利矩阵与反对称张量 ε\varepsilon 之间的两个重要恒等式。在 Weyl 表示中,四维泡利矩阵定义为 σμ=(I,σ)\sigma^\mu = (I, \vec{\sigma})σˉμ=(I,σ)\bar{\sigma}^\mu = (I, -\vec{\sigma}),其中 σ=(σ1,σ2,σ3)\vec{\sigma} = (\sigma^1, \sigma^2, \sigma^3) 为标准泡利矩阵。二维反对称张量约定为 ε12=ε21=1\varepsilon^{12} = -\varepsilon^{21} = 1ε12=ε21=1\varepsilon_{12} = -\varepsilon_{21} = -1

(a) 证明 gμνσμαα˙σνββ˙=2εαβεα˙β˙g^{\mu\nu} \sigma_{\mu}^{\alpha\dot{\alpha}} \sigma_{\nu}^{\beta\dot{\beta}} = 2\varepsilon^{\alpha\beta} \varepsilon^{\dot{\alpha}\dot{\beta}}

展开左边的缩并,利用度规 gμν=diag(1,1,1,1)g^{\mu\nu} = \text{diag}(1, -1, -1, -1),可得: gμνσμαα˙σνββ˙=σ0αα˙σ0ββ˙i=13σiαα˙σiββ˙g^{\mu\nu} \sigma_{\mu}^{\alpha\dot{\alpha}} \sigma_{\nu}^{\beta\dot{\beta}} = \sigma_0^{\alpha\dot{\alpha}} \sigma_0^{\beta\dot{\beta}} - \sum_{i=1}^3 \sigma_i^{\alpha\dot{\alpha}} \sigma_i^{\beta\dot{\beta}} 由于 σ0=I\sigma_0 = I,其矩阵元为 σ0αα˙=δαα˙\sigma_0^{\alpha\dot{\alpha}} = \delta^{\alpha\dot{\alpha}}。对于空间部分的泡利矩阵,利用其完备性关系(即 Fierz 恒等式的一种形式): i=13σiαα˙σiββ˙=2δαβ˙δβα˙δαα˙δββ˙\sum_{i=1}^3 \sigma_i^{\alpha\dot{\alpha}} \sigma_i^{\beta\dot{\beta}} = 2\delta^{\alpha\dot{\beta}}\delta^{\beta\dot{\alpha}} - \delta^{\alpha\dot{\alpha}}\delta^{\beta\dot{\beta}} 将上述两式代入原式中,得到: gμνσμαα˙σνββ˙=δαα˙δββ˙(2δαβ˙δβα˙δαα˙δββ˙)=2(δαα˙δββ˙δαβ˙δβα˙)g^{\mu\nu} \sigma_{\mu}^{\alpha\dot{\alpha}} \sigma_{\nu}^{\beta\dot{\beta}} = \delta^{\alpha\dot{\alpha}}\delta^{\beta\dot{\beta}} - \left( 2\delta^{\alpha\dot{\beta}}\delta^{\beta\dot{\alpha}} - \delta^{\alpha\dot{\alpha}}\delta^{\beta\dot{\beta}} \right) = 2\left( \delta^{\alpha\dot{\alpha}}\delta^{\beta\dot{\beta}} - \delta^{\alpha\dot{\beta}}\delta^{\beta\dot{\alpha}} \right) 另一方面,两个二维反对称张量的乘积可以表示为 Kronecker delta 的行列式: εαβεα˙β˙=det(δαα˙δαβ˙δβα˙δββ˙)=δαα˙δββ˙δαβ˙δβα˙\varepsilon^{\alpha\beta} \varepsilon^{\dot{\alpha}\dot{\beta}} = \det \begin{pmatrix} \delta^{\alpha\dot{\alpha}} & \delta^{\alpha\dot{\beta}} \\ \delta^{\beta\dot{\alpha}} & \delta^{\beta\dot{\beta}} \end{pmatrix} = \delta^{\alpha\dot{\alpha}}\delta^{\beta\dot{\beta}} - \delta^{\alpha\dot{\beta}}\delta^{\beta\dot{\alpha}} 比较两式,即可得出结论: gμνσμαα˙σνββ˙=2εαβεα˙β˙\boxed{ g^{\mu\nu} \sigma_{\mu}^{\alpha\dot{\alpha}} \sigma_{\nu}^{\beta\dot{\beta}} = 2\varepsilon^{\alpha\beta} \varepsilon^{\dot{\alpha}\dot{\beta}} }

(b) 证明 εαβεα˙β˙σμββ˙=σˉα˙αμ\varepsilon_{\alpha\beta} \varepsilon_{\dot{\alpha}\dot{\beta}} \sigma^{\mu\beta\dot{\beta}} = \bar{\sigma}_{\dot{\alpha}\alpha}^{\mu}

我们将左边的表达式写成矩阵乘法的形式。注意到 εα˙β˙=εβ˙α˙\varepsilon_{\dot{\alpha}\dot{\beta}} = -\varepsilon_{\dot{\beta}\dot{\alpha}},可以将其重写为: εαβεα˙β˙σμββ˙=εαβσμββ˙(εβ˙α˙)=(εσμε)αα˙\varepsilon_{\alpha\beta} \varepsilon_{\dot{\alpha}\dot{\beta}} \sigma^{\mu\beta\dot{\beta}} = \varepsilon_{\alpha\beta} \sigma^{\mu\beta\dot{\beta}} (-\varepsilon_{\dot{\beta}\dot{\alpha}}) = - (\varepsilon \sigma^\mu \varepsilon)_{\alpha\dot{\alpha}} 由于 εT=ε\varepsilon^T = -\varepsilon,上式等价于矩阵 (εσμεT)(\varepsilon \sigma^\mu \varepsilon^T)αα˙\alpha\dot{\alpha} 矩阵元。在矩阵表示下,ε=iσ2=(0110)\varepsilon = -i\sigma^2 = \begin{pmatrix} 0 & -1 \\ 1 & 0 \end{pmatrix}

分量讨论如下:

  1. 对于 μ=0\mu = 0σ0=I\sigma^0 = Iεσ0εT=εIεT=εεT=I\varepsilon \sigma^0 \varepsilon^T = \varepsilon I \varepsilon^T = \varepsilon \varepsilon^T = I 因此,εαβεα˙β˙σ0ββ˙=Iαα˙=σˉα˙α0\varepsilon_{\alpha\beta} \varepsilon_{\dot{\alpha}\dot{\beta}} \sigma^{0\beta\dot{\beta}} = I_{\alpha\dot{\alpha}} = \bar{\sigma}^0_{\dot{\alpha}\alpha}

  2. 对于 μ=i{1,2,3}\mu = i \in \{1, 2, 3\}σi\sigma^i 为泡利矩阵: εσiεT=(iσ2)σi(iσ2)=σ2σiσ2\varepsilon \sigma^i \varepsilon^T = (-i\sigma^2) \sigma^i (i\sigma^2) = \sigma^2 \sigma^i \sigma^2 利用泡利矩阵的代数性质 σ2σ1σ2=σ1=(σ1)T\sigma^2 \sigma^1 \sigma^2 = -\sigma^1 = -(\sigma^1)^Tσ2σ2σ2=σ2=(σ2)T\sigma^2 \sigma^2 \sigma^2 = \sigma^2 = -(\sigma^2)^Tσ2σ3σ2=σ3=(σ3)T\sigma^2 \sigma^3 \sigma^2 = -\sigma^3 = -(\sigma^3)^T,可以统一写为: εσiεT=(σi)T\varepsilon \sigma^i \varepsilon^T = -(\sigma^i)^T 取其 αα˙\alpha\dot{\alpha} 矩阵元,并利用转置的定义 (AT)αα˙=Aα˙α(A^T)_{\alpha\dot{\alpha}} = A_{\dot{\alpha}\alpha},得到: εαβεα˙β˙σiββ˙=(σi)αα˙T=σα˙αi\varepsilon_{\alpha\beta} \varepsilon_{\dot{\alpha}\dot{\beta}} \sigma^{i\beta\dot{\beta}} = -(\sigma^i)^T_{\alpha\dot{\alpha}} = -\sigma^i_{\dot{\alpha}\alpha} 根据定义 σˉi=σi\bar{\sigma}^i = -\sigma^i,上式即为 σˉα˙αi\bar{\sigma}^i_{\dot{\alpha}\alpha}

综合 μ=0\mu=0μ=i\mu=i 的情况,对所有 μ\mu 均有: εαβεα˙β˙σμββ˙=σˉα˙αμ\boxed{ \varepsilon_{\alpha\beta} \varepsilon_{\dot{\alpha}\dot{\beta}} \sigma^{\mu\beta\dot{\beta}} = \bar{\sigma}_{\dot{\alpha}\alpha}^{\mu} }

10.4

Problem 10.4

schwarzChapter 10

习题 10.4

来源: 第10章, PDF第182页


10.4 Majorana representation.

(a) Write out the form of the Lorentz generators in the Majorana representation.

(b) Compute J2\vec{J}^2 in the Majorana representation, the left-handed Weyl representation and 4-vector representation. How do you interpret the eigenvalues of J2\vec{J}^2?

(c) Calculate γ5=iγ0γ1γ2γ3\gamma^5 = i\gamma^0 \gamma^1 \gamma^2 \gamma^3 in the Majorana representation.

习题 10.4 - 解答


习题 10.4 分析与解答

(a) Majorana表象下的洛伦兹生成元

在Majorana表象中,所有的Dirac γ\gamma 矩阵都被选取为纯虚数,从而使得Dirac方程可以取实数形式。一种标准的Majorana表象定义如下(利用泡利矩阵 σi\sigma^i):

γ0=(0σ2σ20),γ1=(iσ300iσ3),γ2=(0σ2σ20),γ3=(iσ100iσ1)\gamma^0 = \begin{pmatrix} 0 & \sigma^2 \\ \sigma^2 & 0 \end{pmatrix}, \quad \gamma^1 = \begin{pmatrix} i\sigma^3 & 0 \\ 0 & i\sigma^3 \end{pmatrix}, \quad \gamma^2 = \begin{pmatrix} 0 & -\sigma^2 \\ \sigma^2 & 0 \end{pmatrix}, \quad \gamma^3 = \begin{pmatrix} -i\sigma^1 & 0 \\ 0 & -i\sigma^1 \end{pmatrix}

由于 σ2\sigma^2 是纯虚数,而 σ1,σ3\sigma^1, \sigma^3 是实数,上述所有 γμ\gamma^\mu 均为纯虚数矩阵,且满足Clifford代数 {γμ,γν}=2ημν\{\gamma^\mu, \gamma^\nu\} = 2\eta^{\mu\nu}

洛伦兹群的生成元定义为 Sμν=i4[γμ,γν]S^{\mu\nu} = \frac{i}{4}[\gamma^\mu, \gamma^\nu]。对于 μν\mu \neq \nu,有 Sμν=i2γμγνS^{\mu\nu} = \frac{i}{2}\gamma^\mu\gamma^\nu。 空间旋转生成元 Ji=12ϵijkSjkJ^i = \frac{1}{2}\epsilon^{ijk}S^{jk}

J1=S23=i2γ2γ3=i2(0iσ2σ1iσ2σ10)=i2(0σ3σ30)J^1 = S^{23} = \frac{i}{2}\gamma^2\gamma^3 = \frac{i}{2} \begin{pmatrix} 0 & i\sigma^2\sigma^1 \\ -i\sigma^2\sigma^1 & 0 \end{pmatrix} = \frac{i}{2} \begin{pmatrix} 0 & \sigma^3 \\ -\sigma^3 & 0 \end{pmatrix}
J2=S31=i2γ3γ1=i2(σ1σ300σ1σ3)=12(σ200σ2)J^2 = S^{31} = \frac{i}{2}\gamma^3\gamma^1 = \frac{i}{2} \begin{pmatrix} \sigma^1\sigma^3 & 0 \\ 0 & \sigma^1\sigma^3 \end{pmatrix} = \frac{1}{2} \begin{pmatrix} \sigma^2 & 0 \\ 0 & \sigma^2 \end{pmatrix}
J3=S12=i2γ1γ2=i2(0iσ3σ2iσ3σ20)=i2(0σ1σ10)J^3 = S^{12} = \frac{i}{2}\gamma^1\gamma^2 = \frac{i}{2} \begin{pmatrix} 0 & -i\sigma^3\sigma^2 \\ i\sigma^3\sigma^2 & 0 \end{pmatrix} = \frac{i}{2} \begin{pmatrix} 0 & -\sigma^1 \\ \sigma^1 & 0 \end{pmatrix}

洛伦兹推拨(Boost)生成元 Ki=S0iK^i = S^{0i}

K1=S01=i2γ0γ1=i2(0σ1σ10)K^1 = S^{01} = \frac{i}{2}\gamma^0\gamma^1 = \frac{i}{2} \begin{pmatrix} 0 & -\sigma^1 \\ -\sigma^1 & 0 \end{pmatrix}
K2=S02=i2γ0γ2=i2(I00I)K^2 = S^{02} = \frac{i}{2}\gamma^0\gamma^2 = \frac{i}{2} \begin{pmatrix} I & 0 \\ 0 & -I \end{pmatrix}
K3=S03=i2γ0γ3=i2(0σ3σ30)K^3 = S^{03} = \frac{i}{2}\gamma^0\gamma^3 = \frac{i}{2} \begin{pmatrix} 0 & -\sigma^3 \\ -\sigma^3 & 0 \end{pmatrix}

最终结果为:

Ji=12(δi2σ2iϵi2kσkiϵi2kσkδi2σ2),Ki=i2(δi2Iσ1δi1σ3δi3σ1δi1σ3δi3δi2I)\boxed{ J^i = \frac{1}{2} \begin{pmatrix} \delta_{i2}\sigma^2 & i\epsilon_{i2k}\sigma^k \\ -i\epsilon_{i2k}\sigma^k & \delta_{i2}\sigma^2 \end{pmatrix}, \quad K^i = \frac{i}{2} \begin{pmatrix} \delta_{i2}I & -\sigma^1\delta_{i1}-\sigma^3\delta_{i3} \\ -\sigma^1\delta_{i1}-\sigma^3\delta_{i3} & -\delta_{i2}I \end{pmatrix} }

(b) 计算 J2\vec{J}^2 及其本征值的物理解释

  1. Majorana表象: 利用 (a) 中的结果计算 (Ji)2(J^i)^2
(J1)2=14(σ3σ300σ3σ3)=14I4×4(J^1)^2 = -\frac{1}{4} \begin{pmatrix} -\sigma^3\sigma^3 & 0 \\ 0 & -\sigma^3\sigma^3 \end{pmatrix} = \frac{1}{4} I_{4\times 4}

同理可得 (J2)2=14I4×4(J^2)^2 = \frac{1}{4} I_{4\times 4}(J3)2=14I4×4(J^3)^2 = \frac{1}{4} I_{4\times 4}。因此:

JMajorana2=34I4×4\boxed{ \vec{J}^2_{\text{Majorana}} = \frac{3}{4} I_{4\times 4} }
  1. 左手Weyl表象: 在左手Weyl旋量表示中,旋转生成元直接由泡利矩阵给出 Ji=12σiJ^i = \frac{1}{2}\sigma^i
JWeyl2=i=13(12σi)2=34I2×2\boxed{ \vec{J}^2_{\text{Weyl}} = \sum_{i=1}^3 \left(\frac{1}{2}\sigma^i\right)^2 = \frac{3}{4} I_{2\times 2} }
  1. 4-vector表象: 在四维矢量表示中,旋转生成元的矩阵元为 (Ji) νμ=iϵ  νiμ(J^i)^\mu_{\ \nu} = -i \epsilon^{i\mu}_{\ \ \nu}(仅空间分量非零,即 μ,ν{1,2,3}\mu,\nu \in \{1,2,3\})。 计算其平方和,作用在时间分量上为0,作用在空间分量上为 1(1+1)=21(1+1)=2
J4-vector2=diag(0,2,2,2)\boxed{ \vec{J}^2_{\text{4-vector}} = \text{diag}(0, 2, 2, 2) }

物理解释J2\vec{J}^2 是空间旋转群 SU(2)SU(2) 的Casimir算符,其本征值对应于表示的自旋 j(j+1)j(j+1)

  • Majorana和Weyl表象描述的是自旋 1/21/2 的费米子,因此其本征值均为 12(12+1)=34\frac{1}{2}(\frac{1}{2}+1) = \frac{3}{4}
  • 4-vector表象描述的是时空矢量。在空间旋转下,四维矢量分解为一个标量(时间分量,自旋 j=0j=0,本征值为 00)和一个三维空间矢量(空间分量,自旋 j=1j=1,本征值为 1(1+1)=21(1+1)=2)。

(c) 在Majorana表象中计算 γ5\gamma^5

根据定义 γ5=iγ0γ1γ2γ3\gamma^5 = i\gamma^0 \gamma^1 \gamma^2 \gamma^3。我们分步相乘: 首先计算 γ0γ1\gamma^0 \gamma^1γ2γ3\gamma^2 \gamma^3

γ0γ1=(0σ2σ20)(iσ300iσ3)=(0iσ2σ3iσ2σ30)=(0σ1σ10)\gamma^0 \gamma^1 = \begin{pmatrix} 0 & \sigma^2 \\ \sigma^2 & 0 \end{pmatrix} \begin{pmatrix} i\sigma^3 & 0 \\ 0 & i\sigma^3 \end{pmatrix} = \begin{pmatrix} 0 & i\sigma^2\sigma^3 \\ i\sigma^2\sigma^3 & 0 \end{pmatrix} = \begin{pmatrix} 0 & -\sigma^1 \\ -\sigma^1 & 0 \end{pmatrix}
γ2γ3=(0σ2σ20)(iσ100iσ1)=(0iσ2σ1iσ2σ10)=(0σ3σ30)\gamma^2 \gamma^3 = \begin{pmatrix} 0 & -\sigma^2 \\ \sigma^2 & 0 \end{pmatrix} \begin{pmatrix} -i\sigma^1 & 0 \\ 0 & -i\sigma^1 \end{pmatrix} = \begin{pmatrix} 0 & i\sigma^2\sigma^1 \\ -i\sigma^2\sigma^1 & 0 \end{pmatrix} = \begin{pmatrix} 0 & \sigma^3 \\ -\sigma^3 & 0 \end{pmatrix}

将两者相乘:

γ0γ1γ2γ3=(0σ1σ10)(0σ3σ30)=(σ1σ300σ1σ3)=(iσ200iσ2)\gamma^0 \gamma^1 \gamma^2 \gamma^3 = \begin{pmatrix} 0 & -\sigma^1 \\ -\sigma^1 & 0 \end{pmatrix} \begin{pmatrix} 0 & \sigma^3 \\ -\sigma^3 & 0 \end{pmatrix} = \begin{pmatrix} \sigma^1\sigma^3 & 0 \\ 0 & -\sigma^1\sigma^3 \end{pmatrix} = \begin{pmatrix} -i\sigma^2 & 0 \\ 0 & i\sigma^2 \end{pmatrix}

最后乘以 ii

γ5=i(iσ200iσ2)=(σ200σ2)\gamma^5 = i \begin{pmatrix} -i\sigma^2 & 0 \\ 0 & i\sigma^2 \end{pmatrix} = \begin{pmatrix} \sigma^2 & 0 \\ 0 & -\sigma^2 \end{pmatrix}

因此,Majorana表象下的 γ5\gamma^5 为:

γ5=(σ200σ2)\boxed{ \gamma^5 = \begin{pmatrix} \sigma^2 & 0 \\ 0 & -\sigma^2 \end{pmatrix} }

(注:由于 σ2\sigma^2 是纯虚数,γ5\gamma^5 在Majorana表象中同样是纯虚数矩阵,这保证了Majorana旋量条件的自洽性。)

10.5

Problem 10.5

schwarzChapter 10

习题 10.5

来源: 第10章, PDF第182,183页


10.5 Supersymmetry.

(a) Show that the Lagrangian

L=μϕμϕ+χiσˉχ+FF+mϕF+i2mχTσ2χ+h.c.(10.143)\mathcal{L} = \partial_{\mu} \phi^* \partial^{\mu} \phi + \chi^{\dagger} i \bar{\sigma} \partial \chi + F^* F + m \phi F + \frac{i}{2} m \chi^T \sigma^2 \chi + h.c. \tag{10.143}

is invariant under

δϕ=iϵTσ2χ,δχ=ϵF+σμμϕσ2ϵ,δF=iϵσˉμμχ,\begin{align} \delta \phi &= -i \epsilon^T \sigma^2 \chi, \tag{10.144} \\ \delta \chi &= \epsilon F + \sigma^{\mu} \partial_{\mu} \phi \sigma^2 \epsilon^*, \tag{10.145} \\ \delta F &= -i \epsilon^{\dagger} \bar{\sigma}^{\mu} \partial_{\mu} \chi, \tag{10.146} \\ \end{align}

where ϵ\epsilon is a spinor, χ\chi is a spinor, and FF and ϕ\phi are scalars. All spinors anticommute. σ2\sigma^2 is the second Pauli spin matrix.

** (b) ** The field FF is an auxiliary field, since it has no kinetic term. A useful trick for dealing with auxiliary fields is to solve their equations of motion exactly and plug the result back into the Lagrangian. This is called integrating out a field. Integrate out FF to show that ϕ\phi and χ\chi have the same mass.

** (c) ** Auxiliary fields such as FF act like Lagrange multipliers. One reason to keep the auxiliary fields in the Lagrangian is because they make symmetry transformations exact at the level of the Lagrangian. After the field has been integrated out, the symmetries are only guaranteed to hold if you use the equations of motion. Still using δϕ=iϵTσ2χ\delta \phi = i \epsilon^T \sigma^2 \chi, what is the transformation of χ\chi that makes the Lagrangian in (b) invariant, if you are allowed to use the equations of motion?

习题 10.5 - 解答


(a) 证明拉格朗日量在超对称变换下的不变性

先分析拉格朗日量 L\mathcal{L} 在给定超对称变换下的变分 δL\delta \mathcal{L}。为了证明其不变性,需要证明 δL\delta \mathcal{L} 是一个全导数项。将拉格朗日量分为动能项 Lkin\mathcal{L}_{kin} 和质量项 Lmass\mathcal{L}_{mass} 分别处理。

1. 动能项的变分 动能项为 Lkin=μϕμϕ+χiσˉμμχ+FF\mathcal{L}_{kin} = \partial_{\mu} \phi^* \partial^{\mu} \phi + \chi^{\dagger} i \bar{\sigma}^{\mu} \partial_{\mu} \chi + F^* F。 分别计算各场的变分(注意旋量反易位关系及 σ2\sigma^2 的反对称性 (σ2)T=σ2(\sigma^2)^T = -\sigma^2): 对于标量场 ϕ\phi

δϕ=iϵTσ2χ    δϕ=(δϕ)=iχσ2ϵ\delta \phi = -i \epsilon^T \sigma^2 \chi \implies \delta \phi^* = (\delta \phi)^\dagger = i \chi^\dagger \sigma^2 \epsilon^*

因此,标量动能项的变分为:

δ(μϕμϕ)=iμ(χσ2ϵ)μϕiμϕϵTσ2μχ\delta(\partial_{\mu} \phi^* \partial^{\mu} \phi) = i \partial_{\mu}(\chi^\dagger \sigma^2 \epsilon^*) \partial^{\mu} \phi - i \partial_{\mu} \phi^* \epsilon^T \sigma^2 \partial^{\mu} \chi

对于费米场 χ\chi

δχ=ϵF+σννϕσ2ϵ    (δχ)=ϵF+ϵTσ2σννϕ\delta \chi = \epsilon F + \sigma^{\nu} \partial_{\nu} \phi \sigma^2 \epsilon^* \implies (\delta \chi)^\dagger = \epsilon^\dagger F^* + \epsilon^T \sigma^2 \sigma^{\nu} \partial_{\nu} \phi^*

费米动能项的变分为:

δ(χiσˉμμχ)=(δχ)iσˉμμχ+χiσˉμμ(δχ)=iFϵσˉμμχ+iϵTσ2σνσˉμνϕμχ+iχσˉμϵμF+iχσˉμσνσ2ϵμνϕ\begin{align} \delta(\chi^{\dagger} i \bar{\sigma}^{\mu} \partial_{\mu} \chi) &= (\delta \chi)^\dagger i \bar{\sigma}^{\mu} \partial_{\mu} \chi + \chi^\dagger i \bar{\sigma}^{\mu} \partial_{\mu} (\delta \chi) \\ &= i F^* \epsilon^\dagger \bar{\sigma}^{\mu} \partial_{\mu} \chi + i \epsilon^T \sigma^2 \sigma^{\nu} \bar{\sigma}^{\mu} \partial_{\nu} \phi^* \partial_{\mu} \chi + i \chi^\dagger \bar{\sigma}^{\mu} \epsilon \partial_{\mu} F + i \chi^\dagger \bar{\sigma}^{\mu} \sigma^{\nu} \sigma^2 \epsilon^* \partial_{\mu} \partial_{\nu} \phi \end{align}

对于辅助场 FF

δF=iϵσˉμμχ    δF=iμχσˉμϵ\delta F = -i \epsilon^\dagger \bar{\sigma}^{\mu} \partial_{\mu} \chi \implies \delta F^* = i \partial_{\mu} \chi^\dagger \bar{\sigma}^{\mu} \epsilon

辅助场项的变分为:

δ(FF)=iμχσˉμϵFiFϵσˉμμχ\delta(F^* F) = i \partial_{\mu} \chi^\dagger \bar{\sigma}^{\mu} \epsilon F - i F^* \epsilon^\dagger \bar{\sigma}^{\mu} \partial_{\mu} \chi

将上述三部分相加,可以发现:

  • 包含 FF^* 的项 iFϵσˉμμχi F^* \epsilon^\dagger \bar{\sigma}^{\mu} \partial_{\mu} \chiiFϵσˉμμχ- i F^* \epsilon^\dagger \bar{\sigma}^{\mu} \partial_{\mu} \chi 相互抵消。
  • 包含 FF 的项合并为全导数:iχσˉμϵμF+iμχσˉμϵF=μ(iχσˉμϵF)i \chi^\dagger \bar{\sigma}^{\mu} \epsilon \partial_{\mu} F + i \partial_{\mu} \chi^\dagger \bar{\sigma}^{\mu} \epsilon F = \partial_{\mu}(i \chi^\dagger \bar{\sigma}^{\mu} \epsilon F)
  • 包含 ϕ\phi 二阶导数的项,利用对称性 μνϕ\partial_{\mu} \partial_{\nu} \phi 和恒等式 σˉμσν+σˉνσμ=2ημν\bar{\sigma}^{\mu} \sigma^{\nu} + \bar{\sigma}^{\nu} \sigma^{\mu} = 2\eta^{\mu\nu},有: iχσˉμσνσ2ϵμνϕ=iχημνσ2ϵμνϕ=iχσ2ϵ2ϕi \chi^\dagger \bar{\sigma}^{\mu} \sigma^{\nu} \sigma^2 \epsilon^* \partial_{\mu} \partial_{\nu} \phi = i \chi^\dagger \eta^{\mu\nu} \sigma^2 \epsilon^* \partial_{\mu} \partial_{\nu} \phi = i \chi^\dagger \sigma^2 \epsilon^* \partial^2 \phi 忽略全导数项(分部积分),它等价于 iμχσ2ϵμϕ-i \partial_{\mu} \chi^\dagger \sigma^2 \epsilon^* \partial^{\mu} \phi,这恰好抵消了 δ(μϕμϕ)\delta(\partial_{\mu} \phi^* \partial^{\mu} \phi) 的第一项。
  • 包含 ϕ\phi^* 的项,利用分部积分和恒等式 σνσˉμ+σμσˉν=2ημν\sigma^{\nu} \bar{\sigma}^{\mu} + \sigma^{\mu} \bar{\sigma}^{\nu} = 2\eta^{\mu\nu}iϵTσ2σνσˉμνϕμχ分部积分iϵTσ2ημνμνϕχ=iϵTσ22ϕχ分部积分iϵTσ2μϕμχi \epsilon^T \sigma^2 \sigma^{\nu} \bar{\sigma}^{\mu} \partial_{\nu} \phi^* \partial_{\mu} \chi \xrightarrow{\text{分部积分}} -i \epsilon^T \sigma^2 \eta^{\mu\nu} \partial_{\mu} \partial_{\nu} \phi^* \chi = -i \epsilon^T \sigma^2 \partial^2 \phi^* \chi \xrightarrow{\text{分部积分}} i \epsilon^T \sigma^2 \partial^{\mu} \phi^* \partial_{\mu} \chi 这恰好抵消了 δ(μϕμϕ)\delta(\partial_{\mu} \phi^* \partial^{\mu} \phi) 的第二项。

因此,动能项的变分仅为一个全导数。

2. 质量项的变分 质量项为 Lmass=mϕF+i2mχTσ2χ+h.c.\mathcal{L}_{mass} = m \phi F + \frac{i}{2} m \chi^T \sigma^2 \chi + h.c.。 计算未取共轭部分的变分:

δ(mϕF)=m(δϕ)F+mϕ(δF)=imϵTσ2χFimϕϵσˉμμχ\delta(m \phi F) = m (\delta \phi) F + m \phi (\delta F) = -i m \epsilon^T \sigma^2 \chi F - i m \phi \epsilon^\dagger \bar{\sigma}^{\mu} \partial_{\mu} \chi

对于费米子质量项,利用 χTσ2χ\chi^T \sigma^2 \chi 的对称性:

δ(i2mχTσ2χ)=im(δχ)Tσ2χ=im(ϵTF+μϕ(σμσ2ϵ)T)σ2χ\delta\left(\frac{i}{2} m \chi^T \sigma^2 \chi\right) = i m (\delta \chi)^T \sigma^2 \chi = i m \left( \epsilon^T F + \partial_{\mu} \phi (\sigma^{\mu} \sigma^2 \epsilon^*)^T \right) \sigma^2 \chi

利用转置恒等式 (σμσ2ϵ)T=ϵ(σ2)T(σμ)T=ϵσ2(σμ)T(\sigma^{\mu} \sigma^2 \epsilon^*)^T = \epsilon^\dagger (\sigma^2)^T (\sigma^{\mu})^T = -\epsilon^\dagger \sigma^2 (\sigma^{\mu})^T,以及泡利矩阵恒等式 σ2(σμ)Tσ2=σˉμ\sigma^2 (\sigma^{\mu})^T \sigma^2 = \bar{\sigma}^{\mu}

imμϕ(ϵσ2(σμ)T)σ2χ=imμϕϵσˉμχi m \partial_{\mu} \phi (-\epsilon^\dagger \sigma^2 (\sigma^{\mu})^T) \sigma^2 \chi = -i m \partial_{\mu} \phi \epsilon^\dagger \bar{\sigma}^{\mu} \chi

所以费米质量项变分为:

δ(i2mχTσ2χ)=imFϵTσ2χimμϕϵσˉμχ\delta\left(\frac{i}{2} m \chi^T \sigma^2 \chi\right) = i m F \epsilon^T \sigma^2 \chi - i m \partial_{\mu} \phi \epsilon^\dagger \bar{\sigma}^{\mu} \chi

δ(mϕF)\delta(m \phi F) 与之相加,包含 FF 的项相互抵消,剩余项合并为:

imϕϵσˉμμχimμϕϵσˉμχ=μ(imϕϵσˉμχ)- i m \phi \epsilon^\dagger \bar{\sigma}^{\mu} \partial_{\mu} \chi - i m \partial_{\mu} \phi \epsilon^\dagger \bar{\sigma}^{\mu} \chi = \partial_{\mu} (-i m \phi \epsilon^\dagger \bar{\sigma}^{\mu} \chi)

这也是一个全导数。其厄米共轭 (h.c.) 部分同理也是全导数。

综上所述,δL=μKμ\delta \mathcal{L} = \partial_{\mu} K^{\mu},拉格朗日量在给定的超对称变换下是不变的。


(b) 积分掉辅助场 FF 并证明质量相等

拉格朗日量中包含辅助场 FF 的项为:

LF=FF+mϕF+mϕF\mathcal{L}_F = F^* F + m \phi F + m \phi^* F^*

由于 FF 没有动能项,可以直接通过欧拉-拉格朗日方程求出其运动方程 (EOM):

LF=F+mϕ=0    F=mϕ\frac{\partial \mathcal{L}}{\partial F^*} = F + m \phi^* = 0 \implies F = -m \phi^*
LF=F+mϕ=0    F=mϕ\frac{\partial \mathcal{L}}{\partial F} = F^* + m \phi = 0 \implies F^* = -m \phi

将运动方程代回拉格朗日量中(即“积分掉” FF 场):

LF(mϕ)(mϕ)+mϕ(mϕ)+mϕ(mϕ)=m2ϕϕ\mathcal{L}_F \to (-m \phi)(-m \phi^*) + m \phi (-m \phi^*) + m \phi^* (-m \phi) = -m^2 \phi^* \phi

代入后的在壳 (on-shell) 拉格朗日量变为:

Lonshell=μϕμϕm2ϕϕ+χiσˉμμχ+i2mχTσ2χi2mχσ2χ\mathcal{L}_{on-shell} = \partial_{\mu} \phi^* \partial^{\mu} \phi - m^2 \phi^* \phi + \chi^{\dagger} i \bar{\sigma}^{\mu} \partial_{\mu} \chi + \frac{i}{2} m \chi^T \sigma^2 \chi - \frac{i}{2} m \chi^\dagger \sigma^2 \chi^*

从上式可以读出:

  1. 标量场 ϕ\phi 的质量项为 m2ϕϕ-m^2 \phi^* \phi,对应质量为 mm
  2. 费米场 χ\chi 的马约拉纳 (Majorana) 质量项为 i2mχTσ2χ+h.c.\frac{i}{2} m \chi^T \sigma^2 \chi + h.c.,对应质量也为 mm

因此,积分掉辅助场后,标量场 ϕ\phi 和费米场 χ\chi 具有相同的质量 mm


(c) 使用运动方程后的 χ\chi 变换规则

在 (a) 中,原始的超对称变换为 δϕ=iϵTσ2χ\delta \phi = -i \epsilon^T \sigma^2 \chi。题目 (c) 中给出的条件是“继续使用 δϕ=iϵTσ2χ\delta \phi = i \epsilon^T \sigma^2 \chi”,这与 (a) 相比相差一个全局负号。为了保持拉格朗日量的不变性,χ\chi 的变换规则也必须相应提取一个全局负号,即原始变换变为:

δχnew=ϵFσμμϕσ2ϵ\delta \chi_{new} = -\epsilon F - \sigma^{\mu} \partial_{\mu} \phi \sigma^2 \epsilon^*

(注:如果认为题目 (c) 中的正号仅为印刷错误而沿用 (a) 的 δϕ\delta \phi,则无需加此负号。这里严格按照题目 (c) 给定的 δϕ\delta \phi 符号进行推导)。

现在允许使用运动方程。将 (b) 中求得的辅助场运动方程 F=mϕF = -m \phi^* 代入上述 δχ\delta \chi 的表达式中:

δχ=ϵ(mϕ)σμμϕσ2ϵ\delta \chi = -\epsilon (-m \phi^*) - \sigma^{\mu} \partial_{\mu} \phi \sigma^2 \epsilon^*

化简后得到在壳超对称变换下 χ\chi 的变换规则:

δχ=mϕϵσμμϕσ2ϵ\boxed{ \delta \chi = m \phi^* \epsilon - \sigma^{\mu} \partial_{\mu} \phi \sigma^2 \epsilon^* }

(若沿用 (a) 中的 δϕ=iϵTσ2χ\delta \phi = -i \epsilon^T \sigma^2 \chi,则答案为 δχ=mϕϵ+σμμϕσ2ϵ\delta \chi = -m \phi^* \epsilon + \sigma^{\mu} \partial_{\mu} \phi \sigma^2 \epsilon^*)