本文或含有大量数学公式,请使用电脑或平板电脑浏览以获得最佳阅读体验
前言
这学期选了一门量子计算的课程,按照大名鼎鼎的费曼学习法,我当然是要出一个量子计算的系列啦~尽可能以浅显易懂的语言讲述知识点,读者看明白了,我就真学会了。
这个系列我打算以 Recipe 的形式写,我有我自己的理由,先看看韦氏词典对这个单词的定义:
按照第二个定义来看,我的文章其实是有一点东拼西凑的感觉的。因为我的量子力学学的并不是很好,加上这门课仅仅是作为一个专业选修课来学,所以我不可能深度的、系统性地去讲解量子计算这门学科(毕竟我也在学习阶段)。这个系列的文章我打算以我的老师布置的作业为提纲,围绕着每一道题目给出我自己的解法,同时阐述一下我对该题目相关的知识点的理解,不一定完全对,但是确实是我思考的证明。
我选择的课程指导书是 Michael A. Nielsen 和 Isaac L. Chuang 的 Quantum Computation and Quantum Information (10th Anniversary Edition) ,同时我还会结合老师提供的答案和自己找的资料对解答进行完善。
这个系列里,每一道题目都是一道 dish,同样的菜不同的人可能有不同的做法,你现在所看到的就是我的 Recipe。
Homework I
exercise 1.1
Q:证明算符的迹满足轮换性:T r ( A B C ) = T r ( B C A ) = T r ( C A B ) Tr(ABC)=Tr(BCA)=Tr(CAB) T r ( A BC ) = T r ( BC A ) = T r ( C A B )
A:利用求迹的性质:T r ( X Y ) = T r ( Y X ) Tr(XY) = Tr(YX) T r ( X Y ) = T r ( Y X ) ,设 B C = M BC = M BC = M ,有:
T r ( A M ) = T r ( M A ) (1) Tr(AM) = Tr(MA)
\tag{1}
T r ( A M ) = T r ( M A ) ( 1 )
因此:T r ( A B C ) = T r ( B C A ) Tr(ABC)=Tr(BCA) T r ( A BC ) = T r ( BC A ) ,同理可证 T r ( B C A ) = T r ( C A B ) Tr(BCA)=Tr(CAB) T r ( BC A ) = T r ( C A B )
这一题最重要的地方便是对(1)式的使用,如果我们可以证明(1)式是正确的,那么此题解法不言而喻。求迹操作是对某方阵 的对角项求和,设有 n 阶方阵 A = { a i j } , B = { b i j } A = \{a_{ij}\},B = \{b_{ij}\} A = { a ij } , B = { b ij } ,则:
T r ( A B ) = ∑ k = 1 n ∑ l = 1 n a k l b l k (2) Tr(AB) = \sum_{k=1}^n\sum_{l=1}^na_{kl}b_{lk}
\tag{2}
T r ( A B ) = k = 1 ∑ n l = 1 ∑ n a k l b l k ( 2 )
A B AB A B 交换时,我们也改一下求和的顺序:
T r ( B A ) = ∑ l = 1 n ∑ k = 1 n b l k a k l = ∑ l = 1 n ∑ k = 1 n a k l b l k (3) Tr(BA) = \sum_{l=1}^n\sum_{k=1}^nb_{lk}a_{kl} = \sum_{l=1}^n\sum_{k=1}^na_{kl}b_{lk}
\tag{3}
T r ( B A ) = l = 1 ∑ n k = 1 ∑ n b l k a k l = l = 1 ∑ n k = 1 ∑ n a k l b l k ( 3 )
因为对称性,(2)和(3)式求和项相同,因此 T r ( A B ) = T r ( B A ) Tr(AB) = Tr(BA) T r ( A B ) = T r ( B A )
exercise 1.2
Q:证明任意厄密矩阵A满足:A = ∑ i a i ∣ α i ⟩ ⟨ α i ∣ A = \sum_ia_i\ket{\alpha_i}\bra{\alpha_i} A = ∑ i a i ∣ α i ⟩ ⟨ α i ∣
A:对厄密矩阵A进行对角化分解:A = U D U † A = UDU^\dagger A = U D U † 。厄密矩阵满足本征方程:
A ∣ α i ⟩ = a i ∣ α i ⟩ (4) A\ket{\alpha_i} = a_i\ket{\alpha_i}
\tag{4}
A ∣ α i ⟩ = a i ∣ α i ⟩ ( 4 )
而分解提取出来的:
U = [ ∣ α 1 ⟩ , ∣ α 2 ⟩ , ∣ α 3 ⟩ , ⋯ ] , U † = [ ⟨ α 1 ∣ ⟨ α 2 ∣ ⟨ α 3 ∣ ⋮ ] (5) U = \left[\ket{\alpha_1},\ket{\alpha_2},\ket{\alpha_3},\cdots\right],
U^\dagger = \begin{bmatrix}
\bra{\alpha_1} \\
\bra{\alpha_2} \\
\bra{\alpha_3} \\
\vdots \\
\end{bmatrix}
\tag{5}
U = [ ∣ α 1 ⟩ , ∣ α 2 ⟩ , ∣ α 3 ⟩ , ⋯ ] , U † = ⟨ α 1 ∣ ⟨ α 2 ∣ ⟨ α 3 ∣ ⋮ ( 5 )
有 I = U U † = ∑ i ∣ α i ⟩ ⟨ α i ∣ I = UU^\dagger = \sum_i\ket{\alpha_i}\bra{\alpha_i} I = U U † = ∑ i ∣ α i ⟩ ⟨ α i ∣ ,对(4)式两边同时乘上⟨ α i ∣ \bra{\alpha_i} ⟨ α i ∣ 求和得:
∑ i A ∣ α i ⟩ ⟨ α i ∣ = ∑ i a i ∣ α i ⟩ ⟨ α i ∣ A ( ∑ i ∣ α i ⟩ ⟨ α i ∣ ) = A I = ∑ i a i ∣ α i ⟩ ⟨ α i ∣ (6) \begin{align}
&\sum_iA\ket{\alpha_i}\bra{\alpha_i} = \sum_ia_i\ket{\alpha_i}\bra{\alpha_i}\\
&A(\sum_i\ket{\alpha_i}\bra{\alpha_i}) = A\mathbb{I} = \sum_ia_i\ket{\alpha_i}\bra{\alpha_i}
\end{align}
\tag{6}
i ∑ A ∣ α i ⟩ ⟨ α i ∣ = i ∑ a i ∣ α i ⟩ ⟨ α i ∣ A ( i ∑ ∣ α i ⟩ ⟨ α i ∣ ) = A I = i ∑ a i ∣ α i ⟩ ⟨ α i ∣ ( 6 )
因此,A = ∑ i a i ∣ α i ⟩ ⟨ α i ∣ A = \sum_ia_i\ket{\alpha_i}\bra{\alpha_i} A = ∑ i a i ∣ α i ⟩ ⟨ α i ∣ 得证。
与其说是量子计算题,不如说这题考的是线性代数的知识,牵扯到了本征值和对角化等知识,了解厄密矩阵的性质之后,这题不算难做。
exercise 1.3
Q:证明任意测量算符A的测量结果平均值可以表示为:⟨ A ⟩ = T r ( A ρ ) \lang A \rang = Tr(A\rho) ⟨ A ⟩ = T r ( A ρ )
A:有 ρ = ∣ ψ ⟩ ⟨ ψ ∣ \rho = \ket{\psi}\bra{\psi} ρ = ∣ ψ ⟩ ⟨ ψ ∣ 为密度矩阵,代入求迹公式:
T r ( A ρ ) = T r ( A ∣ ψ ⟩ ⟨ ψ ∣ ) = ∑ i ⟨ i ∣ A ∣ ψ ⟩ ⟨ ψ ∣ i ⟩ = ∑ i ⟨ ψ ∣ i ⟩ ⟨ i ∣ A ∣ ψ ⟩ = ⟨ ψ ∣ A ∣ ψ ⟩ (7) \begin{matrix}
Tr(A\rho)& = & Tr(A\ket{\psi}\bra{\psi}) & = & \sum_i\bra{i}A\ket{\psi}\bra{\psi}i\rang\\
& = & \sum_i\bra{\psi}i\rang\bra{i}A\ket{\psi} & =& \bra{\psi}A\ket{\psi}
\end{matrix}
\tag{7}
T r ( A ρ ) = = T r ( A ∣ ψ ⟩ ⟨ ψ ∣ ) ∑ i ⟨ ψ ∣ i ⟩ ⟨ i ∣ A ∣ ψ ⟩ = = ∑ i ⟨ i ∣ A ∣ ψ ⟩ ⟨ ψ ∣ i ⟩ ⟨ ψ ∣ A ∣ ψ ⟩ ( 7 )
因此:T r ( A ρ ) = ⟨ ψ ∣ A ∣ ψ ⟩ = ⟨ A ⟩ Tr(A\rho) = \bra{\psi}A\ket{\psi} = \lang A \rang T r ( A ρ ) = ⟨ ψ ∣ A ∣ ψ ⟩ = ⟨ A ⟩
本题比较重要的知识点在用求和符号表示的求迹操作:
T r ( A ) = ∑ i ⟨ i ∣ A ∣ i ⟩ (8) Tr(A) = \sum_i\bra{i}A\ket{i}
\tag{8}
T r ( A ) = i ∑ ⟨ i ∣ A ∣ i ⟩ ( 8 )
从(8)式也可以看出求迹操作的本质,即对行列下标相同的对角项求和。下面补充 ρ \rho ρ 为混合态密度矩阵的情况:若 ρ \rho ρ 为混合态 ρ = ∑ k q k ∣ ψ k ⟩ ⟨ ψ k ∣ , ∑ k q k = 1 \rho = \sum_kq_k\ket{\psi_k}\bra{\psi_k},\sum_kq_k = 1 ρ = ∑ k q k ∣ ψ k ⟩ ⟨ ψ k ∣ , ∑ k q k = 1 ,则
A ρ = A ∑ k q k ∣ ψ k ⟩ ⟨ ψ k ∣ T r ( A ρ ) = ∑ i ⟨ i ∣ ( A ∑ k q k ∣ ψ k ⟩ ⟨ ψ k ∣ ) ∣ i ⟩ = ∑ k q k ∑ i ⟨ i ∣ A ∣ ψ k ⟩ ⟨ ψ k ∣ i ⟩ = ∑ k q k ∑ i ⟨ ψ k ∣ i ⟩ ⟨ i ∣ A ∣ ψ k ⟩ = ∑ k q k ⟨ ψ k ∣ A ∣ ψ k ⟩ (9) \begin{align}
A\rho &= A\sum_kq_k\ket{\psi_k}\bra{\psi_k}\\
Tr(A\rho) &= \sum_i\bra{i}(A\sum_kq_k\ket{\psi_k}\bra{\psi_k})\ket{i}\\
&=\sum_kq_k\sum_i\bra{i}A\ket{\psi_k}\bra{\psi_k}i\rang\\
&=\sum_kq_k\sum_i\bra{\psi_k}i\rang\bra{i}A\ket{\psi_k}\\
&=\sum_kq_k\bra{\psi_k}A\ket{\psi_k}
\end{align}
\tag{9}
A ρ T r ( A ρ ) = A k ∑ q k ∣ ψ k ⟩ ⟨ ψ k ∣ = i ∑ ⟨ i ∣ ( A k ∑ q k ∣ ψ k ⟩ ⟨ ψ k ∣ ) ∣ i ⟩ = k ∑ q k i ∑ ⟨ i ∣ A ∣ ψ k ⟩ ⟨ ψ k ∣ i ⟩ = k ∑ q k i ∑ ⟨ ψ k ∣ i ⟩ ⟨ i ∣ A ∣ ψ k ⟩ = k ∑ q k ⟨ ψ k ∣ A ∣ ψ k ⟩ ( 9 )
由(9)式得 T r ( A ρ ) = ∑ k q k ⟨ ψ k ∣ A ∣ ψ k ⟩ = ⟨ A ⟩ Tr(A\rho) = \sum_kq_k\bra{\psi_k}A\ket{\psi_k} = \lang A \rang T r ( A ρ ) = ∑ k q k ⟨ ψ k ∣ A ∣ ψ k ⟩ = ⟨ A ⟩
exercise 1.4
Q:Consider an experiment in which we prepare the state ∣ 0 ⟩ |0\rangle ∣0 ⟩ with the probability ∣ C 0 ∣ 2 |C_0|^2 ∣ C 0 ∣ 2 , and the state ∣ 1 ⟩ |1\rangle ∣1 ⟩ with the probability ∣ C 1 ∣ 2 |C_1|^2 ∣ C 1 ∣ 2 . How to describe this type of quantum state? Compare the differences and similarities between it with the state C 0 ∣ 0 > + C 1 e i θ ∣ 1 ⟩ C_0\left|0\right>+C_1e^{i\theta}\left|1\right\rangle C 0 ∣ 0 ⟩ + C 1 e i θ ∣ 1 ⟩ .
A:该混合态可表示为:∣ ψ ⟩ = C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ \ket{\psi} = C_0\ket{0}+C_1\ket{1} ∣ ψ ⟩ = C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩
C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ C_0\ket{0}+C_1\ket{1} C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ 和 C 0 ∣ 0 > + C 1 e i θ ∣ 1 ⟩ C_0\left|0\right>+C_1e^{i\theta}\left|1\right\rangle C 0 ∣ 0 ⟩ + C 1 e i θ ∣ 1 ⟩ 的区别是在 ∣ 1 ⟩ \ket{1} ∣ 1 ⟩ 处相差一个相位因子 e i θ e^{i\theta} e i θ ,但是这两个态在求密度矩阵时得到的结果不变。
虽然解答很简单,但是本题终于摸到了量子计算的门槛了。本题引入了最基本的两个量子态 ∣ 0 ⟩ \ket{0} ∣ 0 ⟩ 和 ∣ 1 ⟩ \ket{1} ∣ 1 ⟩ ,以及其混合态的表示。注意,混合态是量子计算的独有属性,即一个状态同时具有 ∣ 0 ⟩ \ket{0} ∣ 0 ⟩ 和 ∣ 1 ⟩ \ket{1} ∣ 1 ⟩ 的部分,其中 ∣ C 0 ∣ 2 + ∣ C 1 ∣ 2 = 1 |C_0|^2+|C_1|^2 = 1 ∣ C 0 ∣ 2 + ∣ C 1 ∣ 2 = 1 。对于解答中的命题也很容易得到证明:
ρ 1 = ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) 2 = ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) ∗ = ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) ( C 0 ∗ ⟨ 0 ∣ + C 1 ∗ ⟨ 1 ∣ ) = C 0 2 ∣ 0 ⟩ ⟨ 0 ∣ + C 1 2 ∣ 1 ⟩ ⟨ 1 ∣ + C 0 C 1 ∗ ∣ 0 ⟩ ⟨ 1 ∣ + C 0 ∗ C 1 ∣ 1 ⟩ ⟨ 0 ∣ ρ 2 = ( C 0 ∣ 0 > + C 1 e i θ ∣ 1 ⟩ ) 2 = ( C 0 ∣ 0 > + C 1 e i θ ∣ 1 ⟩ ) ( C 0 ∣ 0 > + C 1 e i θ ∣ 1 ⟩ ) ∗ = ( C 0 ∣ 0 > + C 1 e i θ ∣ 1 ⟩ ) ( C 0 ∗ ⟨ 0 ∣ + C 1 ∗ e − i θ ⟨ 1 ∣ ) = C 0 2 ∣ 0 ⟩ ⟨ 0 ∣ + C 1 2 ∣ 1 ⟩ ⟨ 1 ∣ + C 0 C 1 ∗ ∣ 0 ⟩ ⟨ 1 ∣ + C 0 ∗ C 1 ∣ 1 ⟩ ⟨ 0 ∣ = ρ 1 (10) \begin{align}
\rho_1 &= (C_0\ket{0}+C_1\ket{1})^2 = (C_0\ket{0}+C_1\ket{1})(C_0\ket{0}+C_1\ket{1})^*\\
&=(C_0\ket{0}+C_1\ket{1})(C_0^*\bra{0}+C_1^*\bra{1})\\
&=C_0^2\ket{0}\bra{0}+C_1^2\ket{1}\bra{1}+C_0C_1^*\ket{0}\bra{1}+C_0^*C_1\ket{1}\bra{0}\\
\\
\rho_2&= (C_0\left|0\right>+C_1e^{i\theta}\left|1\right\rangle)^2 = (C_0\left|0\right>+C_1e^{i\theta}\left|1\right\rangle)(C_0\left|0\right>+C_1e^{i\theta}\left|1\right\rangle)^*\\
&= (C_0\left|0\right>+C_1e^{i\theta}\left|1\right\rangle)(C_0^*\bra{0}+C_1^*e^{-i\theta}\bra{1})\\
&= C_0^2\ket{0}\bra{0}+C_1^2\ket{1}\bra{1}+C_0C_1^*\ket{0}\bra{1}+C_0^*C_1\ket{1}\bra{0}\\
&=\rho_1
\end{align}
\tag{10}
ρ 1 ρ 2 = ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) 2 = ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) ∗ = ( C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ) ( C 0 ∗ ⟨ 0 ∣ + C 1 ∗ ⟨ 1 ∣ ) = C 0 2 ∣ 0 ⟩ ⟨ 0 ∣ + C 1 2 ∣ 1 ⟩ ⟨ 1 ∣ + C 0 C 1 ∗ ∣ 0 ⟩ ⟨ 1 ∣ + C 0 ∗ C 1 ∣ 1 ⟩ ⟨ 0 ∣ = ( C 0 ∣ 0 ⟩ + C 1 e i θ ∣ 1 ⟩ ) 2 = ( C 0 ∣ 0 ⟩ + C 1 e i θ ∣ 1 ⟩ ) ( C 0 ∣ 0 ⟩ + C 1 e i θ ∣ 1 ⟩ ) ∗ = ( C 0 ∣ 0 ⟩ + C 1 e i θ ∣ 1 ⟩ ) ( C 0 ∗ ⟨ 0 ∣ + C 1 ∗ e − i θ ⟨ 1 ∣ ) = C 0 2 ∣ 0 ⟩ ⟨ 0 ∣ + C 1 2 ∣ 1 ⟩ ⟨ 1 ∣ + C 0 C 1 ∗ ∣ 0 ⟩ ⟨ 1 ∣ + C 0 ∗ C 1 ∣ 1 ⟩ ⟨ 0 ∣ = ρ 1 ( 10 )
由此可见,一个全局的相位对量子计算结果的概率密度是没有影响的 ,这一点很重要。
exercise 1.5
Q:A density matrix is a representation of statistical mixtures of quantum states.
Let ∣ ψ ⟩ = a ∣ 0 > + b ∣ 1 > |\psi\rangle=a\left|0\right>+b\left|1\right> ∣ ψ ⟩ = a ∣ 0 ⟩ + b ∣ 1 ⟩ be a qubit state, where a , b a,b a , b are complex numbers and
∣ a ∣ 2 + ∣ b ∣ 2 = 1 |a|^2+|b|^2=1 ∣ a ∣ 2 + ∣ b ∣ 2 = 1 . Write down the density matrix ρ = ∣ ψ ⟩ ⟨ ψ ∣ \rho= |\psi\rangle \langle \psi| ρ = ∣ ψ ⟩ ⟨ ψ ∣ in { ∣ 0 ⟩ , ∣ 1 ⟩ } \{\ket{0},\ket{1}\} { ∣ 0 ⟩ , ∣ 1 ⟩ } basis. Calculate the eigenvectors and eigenvalues of ρ \rho ρ .
Let ρ 0 = ∣ 0 ⟩ ⟨ 0 ∣ \rho_0= |0\rangle \langle 0| ρ 0 = ∣0 ⟩ ⟨ 0∣ and ρ 1 = ∣ 1 ⟩ ⟨ 1 ∣ . \rho_1= |1\rangle \langle 1|. ρ 1 = ∣1 ⟩ ⟨ 1∣. Write down the density matrix σ = ρ 0 + ρ 1 2 \sigma=\frac{\rho_0+\rho_1}2 σ = 2 ρ 0 + ρ 1 in { ∣ 0 ⟩ , ∣ 1 ⟩ } \{|0\rangle,|1\rangle\} { ∣0 ⟩ , ∣1 ⟩} basis. Calculate the eigenvectors and eigenvalues of σ \sigma σ .
Compute T r ( ρ 2 ) Tr(\rho^2) T r ( ρ 2 ) and T r ( σ 2 ) Tr(\sigma^2) T r ( σ 2 ) .
A:
由密度矩阵定义:
ρ = ∣ ψ ⟩ ⟨ ψ ∣ = ( a ∣ 0 ⟩ + b ∣ 1 ⟩ ) ( a ∣ 0 ⟩ + b ∣ 1 ⟩ ) ∗ = ( a ∣ 0 ⟩ + b ∣ 1 ⟩ ) ( a ∗ ⟨ 0 ∣ + b ∗ ⟨ 1 ∣ ) = a 2 ∣ 0 ⟩ ⟨ 0 ∣ + b 2 ∣ 1 ⟩ ⟨ 1 ∣ + a b ∗ ∣ 0 ⟩ ⟨ 1 ∣ + a ∗ b ∣ 1 ⟩ ⟨ 0 ∣ (11) \begin{align}
\rho &= \ket{\psi}\bra{\psi} = (a\ket{0}+b\ket{1})(a\ket{0}+b\ket{1})^*\\
&=(a\ket{0}+b\ket{1})(a^*\bra{0}+b^*\bra{1})\\
&=a^2\ket{0}\bra{0}+b^2\ket{1}\bra{1}+ab^*\ket{0}\bra{1}+a^*b\ket{1}\bra{0}
\end{align}
\tag{11}
ρ = ∣ ψ ⟩ ⟨ ψ ∣ = ( a ∣ 0 ⟩ + b ∣ 1 ⟩ ) ( a ∣ 0 ⟩ + b ∣ 1 ⟩ ) ∗ = ( a ∣ 0 ⟩ + b ∣ 1 ⟩ ) ( a ∗ ⟨ 0 ∣ + b ∗ ⟨ 1 ∣ ) = a 2 ∣ 0 ⟩ ⟨ 0 ∣ + b 2 ∣ 1 ⟩ ⟨ 1 ∣ + a b ∗ ∣ 0 ⟩ ⟨ 1 ∣ + a ∗ b ∣ 1 ⟩ ⟨ 0 ∣ ( 11 )
因为 ∣ 0 ⟩ = [ 1 0 ] , ∣ 1 ⟩ = [ 0 1 ] \ket{0}=\begin{bmatrix}1\\0\end{bmatrix},\ket{1}=\begin{bmatrix}0\\1\end{bmatrix} ∣ 0 ⟩ = [ 1 0 ] , ∣ 1 ⟩ = [ 0 1 ] ,则:
ρ = ∣ ψ ⟩ ⟨ ψ ∣ = [ a 2 a b ∗ a ∗ b b 2 ] (12) \rho = \ket{\psi}\bra{\psi} =
\begin{bmatrix}
a^2 &ab^*\\
a^*b &b^2
\end{bmatrix}
\tag{12}
ρ = ∣ ψ ⟩ ⟨ ψ ∣ = [ a 2 a ∗ b a b ∗ b 2 ] ( 12 )
设本征值为 λ \lambda λ 本征向量为 v \mathbf{v} v ,有 ρ v = λ v → ( ρ − λ I ) v = 0 \rho\mathbf{v} = \lambda\mathbf{v}\rightarrow(\rho-\lambda\mathbb{I})\mathbf{v}=0 ρ v = λ v → ( ρ − λ I ) v = 0 ,则:
∣ a 2 − λ a b ∗ a ∗ b b 2 − λ ∣ = 0 (12) \left|\begin{matrix}
a^2 - \lambda &ab^*\\
a^*b &b^2 - \lambda
\end{matrix}\right| = 0
\tag{12}
a 2 − λ a ∗ b a b ∗ b 2 − λ = 0 ( 12 )
解得本征值 λ 1 = a 2 + b 2 = 1 , λ 2 = 0 \lambda_1 = a^2+b^2 = 1,\lambda_2 = 0 λ 1 = a 2 + b 2 = 1 , λ 2 = 0 ,代入求得本征向量 v 1 = [ a b ] , v 2 = [ b ∗ − a ∗ ] \mathbf{v_1} = \begin{bmatrix}a\\b\end{bmatrix},\mathbf{v_2} = \begin{bmatrix}b^*\\-a^*\end{bmatrix} v 1 = [ a b ] , v 2 = [ b ∗ − a ∗ ]
由题目得 σ = ρ 0 + ρ 1 2 = 1 2 ( ∣ 0 ⟩ ⟨ 0 ∣ + ∣ 1 ⟩ ⟨ 1 ∣ ) \sigma = \frac{\rho_0+\rho_1}{2} = \frac{1}{2}(|0\rangle \langle 0|+|1\rangle \langle 1|) σ = 2 ρ 0 + ρ 1 = 2 1 ( ∣0 ⟩ ⟨ 0∣ + ∣1 ⟩ ⟨ 1∣ )
σ = 1 2 ( [ 1 0 0 0 ] + [ 0 0 0 1 ] ) = [ 1 2 0 0 1 2 ] (13) \sigma = \frac12(\begin{bmatrix}1 &0\\0 &0\end{bmatrix}+\begin{bmatrix}0 &0\\0 &1\end{bmatrix}) = \begin{bmatrix}\frac12 &0\\0 &\frac12\end{bmatrix}
\tag{13}
σ = 2 1 ( [ 1 0 0 0 ] + [ 0 0 0 1 ] ) = [ 2 1 0 0 2 1 ] ( 13 )
由 σ v = λ v → ( σ − λ I ) v = 0 \sigma\mathbf{v} = \lambda\mathbf{v}\rightarrow(\sigma-\lambda\mathbb{I})\mathbf{v}=0 σ v = λ v → ( σ − λ I ) v = 0 求得简并本征值 λ 1 = λ 2 = 1 2 \lambda_1 = \lambda_2 = \frac12 λ 1 = λ 2 = 2 1 ,本征向量 v 1 = [ 1 0 ] , v 2 = [ 0 1 ] \mathbf{v_1} = \begin{bmatrix}1\\0\end{bmatrix},\mathbf{v_2} = \begin{bmatrix}0\\1\end{bmatrix} v 1 = [ 1 0 ] , v 2 = [ 0 1 ]
由上面两个题目的结果:
T r ( ρ 2 ) = T r ( [ a 2 a b ∗ a ∗ b b 2 ] 2 ) = a 4 + 2 a 2 b 2 + b 4 = ( a 2 + b 2 ) 2 = 1 T r ( σ 2 ) = T r ( [ 1 2 0 0 1 2 ] 2 ) = 1 2 (14) \begin{align}
Tr(\rho^2) &=Tr(\begin{bmatrix}
a^2 &ab^*\\
a^*b &b^2
\end{bmatrix}^2) = a^4+2a^2b^2+b^4 = (a^2+b^2)^2 = 1\\
Tr(\sigma^2) &= Tr(\begin{bmatrix}
\frac12 &0\\
0 &\frac12
\end{bmatrix}^2) = \frac12
\end{align}
\tag{14}
T r ( ρ 2 ) T r ( σ 2 ) = T r ( [ a 2 a ∗ b a b ∗ b 2 ] 2 ) = a 4 + 2 a 2 b 2 + b 4 = ( a 2 + b 2 ) 2 = 1 = T r ( [ 2 1 0 0 2 1 ] 2 ) = 2 1 ( 14 )
这道题题目虽多,但是只是帮我们复习了一遍矩阵本征值和本征向量的求法,具体本征值和本征向量的物理意义要在 HW2 结合量子比特的几何图像来进行说明
Homework II
exercise 2.1
Q:Prove that for any 2-dimension linear operator A A A ,
A = 1 2 T r ( A ) I + 1 2 ∑ k = 1 3 T r ( A σ k ) σ k (15) A=\frac{1}{2}Tr(A)\mathbb{I}+\frac{1}{2}\sum_{k=1}^{3}Tr(A\sigma_{k})\sigma_{k}
\tag{15}
A = 2 1 T r ( A ) I + 2 1 k = 1 ∑ 3 T r ( A σ k ) σ k ( 15 )
in which σ k \sigma_k σ k are Pauli matrices.
A:将 A 用 { I , σ 1 , σ 2 , σ 3 } \{\mathbb{I},\sigma_1,\sigma_2,\sigma_3\} { I , σ 1 , σ 2 , σ 3 } 展开:A = α 0 I + α 1 σ 1 + α 2 σ 2 + α 3 σ 3 A = \alpha_0\mathbb{I}+\alpha_1\sigma_1+\alpha_2\sigma_2+\alpha_3\sigma_3 A = α 0 I + α 1 σ 1 + α 2 σ 2 + α 3 σ 3 则
T r ( A ) = T r ( α 0 I + α 1 σ 1 + α 2 σ 2 + α 3 σ 3 ) = T r ( α 0 I ) + T r ( α 1 σ 1 ) + T r ( α 2 σ 2 ) + T r ( α 3 σ 3 ) = 2 α 0 (16) \begin{align}
Tr(A) &= Tr(\alpha_0\mathbb{I}+\alpha_1\sigma_1+\alpha_2\sigma_2+\alpha_3\sigma_3)\\
&=Tr(\alpha_0\mathbb{I})+Tr(\alpha_1\sigma_1)+Tr(\alpha_2\sigma_2)+Tr(\alpha_3\sigma_3)\\
&= 2\alpha_0
\end{align}
\tag{16}
T r ( A ) = T r ( α 0 I + α 1 σ 1 + α 2 σ 2 + α 3 σ 3 ) = T r ( α 0 I ) + T r ( α 1 σ 1 ) + T r ( α 2 σ 2 ) + T r ( α 3 σ 3 ) = 2 α 0 ( 16 )
则 1 2 T r ( A ) = α 0 \frac{1}{2}Tr(A) = \alpha_0 2 1 T r ( A ) = α 0 。展开式两侧同时乘 σ 1 \sigma_1 σ 1 得:A σ 1 = α 0 σ 1 + α 1 I − i α 2 σ 3 + i α 3 σ 2 A\sigma_1 = \alpha_0\sigma_1+\alpha_1\mathbb{I}-i\alpha_2\sigma_3+i\alpha_3\sigma_2 A σ 1 = α 0 σ 1 + α 1 I − i α 2 σ 3 + i α 3 σ 2 ,则 T r ( A σ 1 ) = 2 α 1 Tr(A\sigma_1) = 2\alpha_1 T r ( A σ 1 ) = 2 α 1 。同理,T r ( A σ 2 ) = 2 α 2 , T r ( A σ 3 ) = 2 α 3 Tr(A\sigma_2) = 2\alpha_2,Tr(A\sigma_3) = 2\alpha_3 T r ( A σ 2 ) = 2 α 2 , T r ( A σ 3 ) = 2 α 3 ,代入(15)式得:
1 2 T r ( A ) I + 1 2 ∑ k = 1 3 T r ( A σ k ) σ k = α 0 I + α 1 σ 1 + α 2 σ 2 + α 3 σ 3 = A (17) \frac{1}{2}Tr(A)\mathbb{I}+\frac{1}{2}\sum_{k=1}^{3}Tr(A\sigma_{k})\sigma_{k}
= \alpha_0\mathbb{I}+\alpha_1\sigma_1+\alpha_2\sigma_2+\alpha_3\sigma_3 = A
\tag{17}
2 1 T r ( A ) I + 2 1 k = 1 ∑ 3 T r ( A σ k ) σ k = α 0 I + α 1 σ 1 + α 2 σ 2 + α 3 σ 3 = A ( 17 )
本题提到了一组很重要的矩阵——Pauli 矩阵 σ ⃗ = σ ( σ 1 , σ 2 , σ 3 ) o r ( σ x , σ y , σ z ) \vec{\sigma} = \sigma(\sigma_1,\sigma_2,\sigma_3)or(\sigma_x,\sigma_y,\sigma_z) σ = σ ( σ 1 , σ 2 , σ 3 ) or ( σ x , σ y , σ z )
σ x = [ 0 1 1 0 ] , σ y = [ 0 − i i 0 ] , σ z = [ 1 0 0 − 1 ] (18) \sigma_x = \begin{bmatrix}0 & 1\\1 &0\end{bmatrix},
\sigma_y = \begin{bmatrix}0 & -i\\i &0\end{bmatrix},
\sigma_z = \begin{bmatrix}1 & 0\\0 &-1\end{bmatrix}
\tag{18}
σ x = [ 0 1 1 0 ] , σ y = [ 0 i − i 0 ] , σ z = [ 1 0 0 − 1 ] ( 18 )
由于单个量子比特(qbit)的密度矩阵是一个 2 × 2 2\times2 2 × 2 的矩阵,所以我们经常用一组二维希尔伯特空间的基 { I , σ 1 , σ 2 , σ 3 } \{\mathbb{I},\sigma_1,\sigma_2,\sigma_3\} { I , σ 1 , σ 2 , σ 3 } 来表示它。
一般情况下,我们可以用 n ⃗ = ( x , y , z ) \vec{n} = (x,y,z) n = ( x , y , z ) 将任意密度矩阵转化为下面这个表达方式:
ρ = 1 2 ( I + n ⃗ ⋅ σ ⃗ ) = 1 2 [ 1 + z x − i y x + i y 1 − z ] (19) \rho = \frac12(\mathbb{I}+\vec{n}\cdot\vec{\sigma}) = \frac12\begin{bmatrix}1+z & x-iy\\x+iy &1-z\end{bmatrix}
\tag{19}
ρ = 2 1 ( I + n ⋅ σ ) = 2 1 [ 1 + z x + i y x − i y 1 − z ] ( 19 )
由于密度矩阵的半正定性,有 ∣ n ⃗ ∣ 2 = x 2 + y 2 + z 2 ≤ 1 \mid{\vec{n}}\mid^{2}=x^{2}+y^{2}+z^{2}\leq1 ∣ n ∣ 2 = x 2 + y 2 + z 2 ≤ 1 ,纯态时取等于号。
补充 Pauli 矩阵的计算性质:
σ k σ l = δ k l I + i ∑ m ε k l m σ m T r ( σ k σ l ) = 2 δ k l (20) \begin{align}
&\sigma_k\sigma_l = \delta_{kl}\mathbb{I}+i\sum_m\varepsilon_{klm}\sigma_m\\
&Tr(\sigma_k\sigma_l) = 2\delta_{kl}
\end{align}
\tag{20}
σ k σ l = δ k l I + i m ∑ ε k l m σ m T r ( σ k σ l ) = 2 δ k l ( 20 )
exercise 2.2
Q:Let ρ \rho ρ be a density operator.
Show that ρ \rho ρ can be written as
ρ = I + r ⋅ σ ⃗ 2 (21) \rho=\frac{\mathbb{I}+\mathbf{r}\cdot\vec{\sigma}}2
\tag{21}
ρ = 2 I + r ⋅ σ ( 21 )
where r \boldsymbol{r} r is a real three-dimensional vector and ∣ ∣ r ∣ ∣ ≤ 1. ||\mathbf{r}||\leq1. ∣∣ r ∣∣ ≤ 1.
Show that $Tr( \rho^2) \leq 1, $with equality if and only if ρ \rho ρ is a pure state.
Show that a state ρ \rho ρ is a pure state if and only if ∣ ∣ r ∣ ∣ = 1. ||\boldsymbol{r}||=1. ∣∣ r ∣∣ = 1.
A:
利用 exercise 2.1 的结论 x = T r ( A σ 1 ) = 2 α 1 , y = T r ( A σ 2 ) = 2 α 2 , z = T r ( A σ 3 ) = 2 α 3 x = Tr(A\sigma_1) = 2\alpha_1,y = Tr(A\sigma_2) = 2\alpha_2,z = Tr(A\sigma_3) = 2\alpha_3 x = T r ( A σ 1 ) = 2 α 1 , y = T r ( A σ 2 ) = 2 α 2 , z = T r ( A σ 3 ) = 2 α 3 和 A = 1 2 T r ( A ) I + 1 2 ∑ k = 1 3 T r ( A σ k ) σ k A=\frac{1}{2}Tr(A)\mathbb{I}+\frac{1}{2}\sum_{k=1}^{3}Tr(A\sigma_{k})\sigma_{k} A = 2 1 T r ( A ) I + 2 1 ∑ k = 1 3 T r ( A σ k ) σ k ,有 ρ = 1 2 ( I + x σ 1 + y σ 2 + z σ 3 ) \rho = \frac12(\mathbb{I}+x\sigma_1+y\sigma_2+z\sigma_3) ρ = 2 1 ( I + x σ 1 + y σ 2 + z σ 3 ) 。
令 r = ( x , y , z ) \mathbf{r} = (x,y,z) r = ( x , y , z ) ,则 ρ = ( I + r ⋅ σ ⃗ ) / 2 \rho=(\mathbb{I}+\mathbf{r}\cdot\vec{\sigma})/2 ρ = ( I + r ⋅ σ ) /2
设 ρ = ( I + r ⋅ σ ⃗ ) / 2 \rho=(\mathbb{I}+\mathbf{r}\cdot\vec{\sigma})/2 ρ = ( I + r ⋅ σ ) /2
则 ρ 2 = ( I + r ⋅ σ ⃗ ) 2 / 4 = [ I 2 + 2 r ⋅ σ ⃗ + ( r ⋅ σ ⃗ ) 2 ] / 4 \rho^2=(\mathbb{I}+\mathbf{r}\cdot\vec{\sigma})^2/4 = [\mathbb{I}^2+2\mathbf{r}\cdot\vec{\sigma}+(\mathbf{r}\cdot\vec{\sigma})^2]/4 ρ 2 = ( I + r ⋅ σ ) 2 /4 = [ I 2 + 2 r ⋅ σ + ( r ⋅ σ ) 2 ] /4 ,因为有:
( m ⋅ σ ⃗ ) ( n ⋅ σ ⃗ ) = m ⋅ n I + i ( m × n ) ⋅ σ ⃗ (22) (\mathbf{m}\cdot\vec{\sigma})(\mathbf{n}\cdot\vec{\sigma}) = \mathbf{m\cdot n}\mathbb{I}+i(\mathbf{m\times n})\cdot\vec{\sigma}
\tag{22}
( m ⋅ σ ) ( n ⋅ σ ) = m ⋅ n I + i ( m × n ) ⋅ σ ( 22 )
则 ( r ⋅ σ ⃗ ) 2 = r ⋅ r I + i ( r × r ) ⋅ σ ⃗ = ∣ r ∣ 2 I (\mathbf{r}\cdot\vec{\sigma})^2 = \mathbf{r}\cdot\mathbf{r}\mathbb{I}+i(\mathbf{r\times r})\cdot\vec{\sigma} = \mathbf{|r|}^2\mathbb{I} ( r ⋅ σ ) 2 = r ⋅ r I + i ( r × r ) ⋅ σ = ∣r∣ 2 I ,代入原式得:
T r ( ρ 2 ) = 1 4 T r ( I 2 + 2 r ⋅ σ ⃗ + ∣ r ∣ 2 I ) = 1 2 ( 1 + ∣ r ∣ 2 ) (23) \begin{align}
Tr(\rho^2) &= \frac14Tr(\mathbb{I}^2+2\mathbf{r}\cdot\vec{\sigma}+\mathbf{|r|}^2\mathbb{I})\\
&= \frac12(1+\mathbf{|r|}^2)
\end{align}
\tag{23}
T r ( ρ 2 ) = 4 1 T r ( I 2 + 2 r ⋅ σ + ∣r∣ 2 I ) = 2 1 ( 1 + ∣r∣ 2 ) ( 23 )
由于 ∣ r ∣ 2 ≤ 1 \mathbf{|r|}^2\leq1 ∣r∣ 2 ≤ 1 ,所以 T r ( ρ 2 ) ≤ 1 Tr(\rho^2)\leq1 T r ( ρ 2 ) ≤ 1 。
1)当 ρ \rho ρ 为纯态时,ρ = ∣ ψ ⟩ ⟨ ψ ∣ \rho = \ket{\psi}\bra{\psi} ρ = ∣ ψ ⟩ ⟨ ψ ∣ 。易得 T r ( ρ 2 ) = T r ( ∣ ψ ⟩ ⟨ ψ ∣ ψ ⟩ ⟨ ψ ∣ ) = 1 Tr(\rho^2) = Tr(\ket{\psi}\lang\psi\ket{\psi}\bra{\psi}) = 1 T r ( ρ 2 ) = T r ( ∣ ψ ⟩ ⟨ ψ ∣ ψ ⟩ ⟨ ψ ∣ ) = 1
2)当 T r ( ρ 2 ) = 1 2 ( 1 + ∣ r ∣ 2 ) = 1 Tr(\rho^2) = \frac12(1+\mathbf{|r|}^2) = 1 T r ( ρ 2 ) = 2 1 ( 1 + ∣r∣ 2 ) = 1 时,∣ r ∣ = 1 \mathbf{|r|} = 1 ∣r∣ = 1 。有:
r ⋅ σ ⃗ = sin θ cos ϕ σ 1 + sin θ sin ϕ σ 2 + cos θ σ 3 = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ (24) \begin{align}
\mathbf{r}\cdot\vec{\sigma} &= \sin\theta\cos\phi\sigma_1+\sin\theta\sin\phi\sigma_2+\cos\theta\sigma_3\\
&=\ket{r+}\bra{r+} - \ket{r-}\bra{r-}
\end{align}
\tag{24}
r ⋅ σ = sin θ cos ϕ σ 1 + sin θ sin ϕ σ 2 + cos θ σ 3 = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ ( 24 )
且 ∣ r + ⟩ ⟨ r + ∣ + ∣ r − ⟩ ⟨ r − ∣ = I \ket{r+}\bra{r+} + \ket{r-}\bra{r-} = \mathbb{I} ∣ r + ⟩ ⟨ r + ∣ + ∣ r − ⟩ ⟨ r − ∣ = I ,因此把 I \mathbb{I} I 和 r ⋅ σ ⃗ \mathbf{r}\cdot\vec{\sigma} r ⋅ σ 代入 ρ = ( I + r ⋅ σ ⃗ ) / 2 \rho=(\mathbb{I}+\mathbf{r}\cdot\vec{\sigma})/2 ρ = ( I + r ⋅ σ ) /2 得 ρ = ∣ r + ⟩ ⟨ r + ∣ \rho = \ket{r+}\bra{r+} ρ = ∣ r + ⟩ ⟨ r + ∣ 为纯态。综上所述 T r ( ρ 2 ) = 1 Tr(\rho^2)=1 T r ( ρ 2 ) = 1 和 ρ \rho ρ 为纯态互为充要条件 。
由第二问结论得 ∣ r ∣ = 1 |\mathbf{r}| = 1 ∣ r ∣ = 1 时,ρ \rho ρ 为纯态。反过来,当 ρ \rho ρ 为纯态时,T r ( ρ 2 ) = 1 2 ( 1 + ∣ r ∣ 2 ) = 1 Tr(\rho^2) = \frac12(1+|\mathbf{r}|^2) = 1 T r ( ρ 2 ) = 2 1 ( 1 + ∣ r ∣ 2 ) = 1 。由于 ∣ r ∣ > 0 |\mathbf{r}| > 0 ∣ r ∣ > 0 ,因此 ∣ r ∣ |\mathbf{r}| ∣ r ∣ 只能为1,得证。
此题适合和 exercise 2.3 合并在一起讲解,如果有看不懂的地方,请移步 exercise 2.3。
此外这是本题用到的重要公式(22)的证明:
( m ⋅ σ ⃗ ) ( n ⋅ σ ⃗ ) = n x m x + n x m y σ x σ y + n x m z σ x σ z + n y m x σ y σ x + n y m y + n y m z σ y σ z + n z m x σ z σ x + n z m y σ z σ y + n z m z = m ⋅ n + i n x m y σ z − i n x m z σ y − i n y m x σ z + i n y m z σ x + i n z m x σ y − i n z m y σ x = m ⋅ n I + i ( m × n ) ⋅ σ ⃗ (25) \begin{align}
(\mathbf{m}\cdot\vec{\sigma})(\mathbf{n}\cdot\vec{\sigma})=& n_xm_x+n_xm_y\sigma_x\sigma_y+n_xm_z\sigma_x\sigma_z+n_ym_x\sigma_y\sigma_x+n_ym_y+\\
&n_ym_z\sigma_y\sigma_z+n_zm_x\sigma_z\sigma_x+n_zm_y\sigma_z\sigma_y+n_zm_z\\
=&\mathbf{m}\cdot\mathbf{n}+in_xm_y\sigma_z-in_xm_z\sigma_y-in_ym_x\sigma_z+in_ym_z\sigma_x+\\
&in_zm_x\sigma_y-in_zm_y\sigma_x\\
=&\mathbf{m\cdot n}\mathbb{I}+i(\mathbf{m\times n})\cdot\vec{\sigma}
\end{align}
\tag{25}
( m ⋅ σ ) ( n ⋅ σ ) = = = n x m x + n x m y σ x σ y + n x m z σ x σ z + n y m x σ y σ x + n y m y + n y m z σ y σ z + n z m x σ z σ x + n z m y σ z σ y + n z m z m ⋅ n + i n x m y σ z − i n x m z σ y − i n y m x σ z + i n y m z σ x + i n z m x σ y − i n z m y σ x m ⋅ n I + i ( m × n ) ⋅ σ ( 25 )
exercise 2.3
Q:给定单位向量 ( sin θ cos ϕ , sin θ sin ϕ , cos θ ) (\sin\theta\cos\phi,\sin\theta\sin\phi,\cos\theta) ( sin θ cos ϕ , sin θ sin ϕ , cos θ ) 定义 σ n ≡ r ⋅ σ ⃗ \sigma_n\equiv \mathbf{r}\cdot\vec{\sigma} σ n ≡ r ⋅ σ ,试证明其本征分解结果为:
σ n = sin θ cos ϕ σ x + sin θ sin ϕ σ y + cos θ σ z = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ (26) \begin{align}
\sigma_n &= \sin\theta\cos\phi\sigma_x+\sin\theta\sin\phi\sigma_y+\cos\theta\sigma_z\\
&=\ket{r+}\bra{r+} - \ket{r-}\bra{r-}
\end{align}
\tag{26}
σ n = sin θ cos ϕ σ x + sin θ sin ϕ σ y + cos θ σ z = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ ( 26 )
其中:
∣ r + ⟩ = cos θ 2 e − i ϕ 2 ∣ 0 ⟩ + sin θ 2 e i ϕ 2 ∣ 1 ⟩ ∣ r − ⟩ = sin θ 2 e − i ϕ 2 ∣ 0 ⟩ − cos θ 2 e i ϕ 2 ∣ 1 ⟩ (27) \begin{align}
\ket{r+} &= \cos\frac\theta2e^{-i\frac\phi2}\ket{0}+\sin\frac\theta2e^{i\frac\phi2}\ket{1}\\
\ket{r-} &= \sin\frac\theta2e^{-i\frac\phi2}\ket{0}-\cos\frac\theta2e^{i\frac\phi2}\ket{1}
\end{align}
\tag{27}
∣ r + ⟩ ∣ r − ⟩ = cos 2 θ e − i 2 ϕ ∣ 0 ⟩ + sin 2 θ e i 2 ϕ ∣ 1 ⟩ = sin 2 θ e − i 2 ϕ ∣ 0 ⟩ − cos 2 θ e i 2 ϕ ∣ 1 ⟩ ( 27 )
且此向量对应的量子态为纯态。
A:由exercise 2.2(24)式:
r ⋅ σ ⃗ = sin θ cos ϕ σ 1 + sin θ sin ϕ σ 2 + cos θ σ 3 = sin θ cos ϕ [ 0 1 1 0 ] + sin θ sin ϕ [ 0 − i i 0 ] + cos θ [ 1 0 0 − 1 ] = [ cos θ sin θ e − i ϕ sin θ e i ϕ − cos θ ] (28) \begin{align}
\mathbf{r}\cdot\vec{\sigma} &= \sin\theta\cos\phi\sigma_1+\sin\theta\sin\phi\sigma_2+\cos\theta\sigma_3\\
&=\sin\theta\cos\phi\begin{bmatrix}0 & 1\\1 &0\end{bmatrix}+\sin\theta\sin\phi\begin{bmatrix}0 & -i\\i &0\end{bmatrix}+\cos\theta\begin{bmatrix}1 & 0\\0 &-1\end{bmatrix}\\
&=\begin{bmatrix}
\cos\theta & \sin\theta e^{-i\phi}\\
\sin\theta e^{i\phi} &-\cos\theta
\end{bmatrix}
\end{align}
\tag{28}
r ⋅ σ = sin θ cos ϕ σ 1 + sin θ sin ϕ σ 2 + cos θ σ 3 = sin θ cos ϕ [ 0 1 1 0 ] + sin θ sin ϕ [ 0 i − i 0 ] + cos θ [ 1 0 0 − 1 ] = [ cos θ sin θ e i ϕ sin θ e − i ϕ − cos θ ] ( 28 )
设本征值为 λ \lambda λ 本征向量为 v \mathbf{v} v ,得 λ 2 − ( cos 2 θ + sin 2 θ ) = 0 \lambda^2 - (\cos^2\theta+\sin^2\theta) = 0 λ 2 − ( cos 2 θ + sin 2 θ ) = 0 ,即 λ = ± 1 \lambda = \pm1 λ = ± 1 ,代入本征值得到本征向量:
v 1 = [ cos θ 2 sin θ 2 e i ϕ ] = cos θ 2 ∣ 0 ⟩ + sin θ 2 e i ϕ ∣ 1 ⟩ (29) \mathbf{v}_1 = \begin{bmatrix}\cos\frac\theta2\\\sin\frac\theta2e^{i\phi}\end{bmatrix}=\cos\frac\theta2\ket{0}+\sin\frac\theta2e^{i\phi}\ket{1}
\tag{29}
v 1 = [ cos 2 θ sin 2 θ e i ϕ ] = cos 2 θ ∣ 0 ⟩ + sin 2 θ e i ϕ ∣ 1 ⟩ ( 29 )
由 exercise 1.4 得知,我们对(26)式提出一个全局相位 e i ϕ / 2 e^{i\phi/2} e i ϕ /2 不改变其密度矩阵,因此我们可以通过此操作得到 ∣ r + ⟩ \ket{r+} ∣ r + ⟩ ,并同理得到 ∣ r − ⟩ \ket{r-} ∣ r − ⟩ :
∣ r + ⟩ = cos θ 2 e − i ϕ 2 ∣ 0 ⟩ + sin θ 2 e i ϕ 2 ∣ 1 ⟩ ∣ r − ⟩ = sin θ 2 e − i ϕ 2 ∣ 0 ⟩ − cos θ 2 e i ϕ 2 ∣ 1 ⟩ (30) \begin{align}
\ket{r+} &= \cos\frac\theta2e^{-i\frac\phi2}\ket{0}+\sin\frac\theta2e^{i\frac\phi2}\ket{1}\\
\ket{r-} &= \sin\frac\theta2e^{-i\frac\phi2}\ket{0}-\cos\frac\theta2e^{i\frac\phi2}\ket{1}
\end{align}
\tag{30}
∣ r + ⟩ ∣ r − ⟩ = cos 2 θ e − i 2 ϕ ∣ 0 ⟩ + sin 2 θ e i 2 ϕ ∣ 1 ⟩ = sin 2 θ e − i 2 ϕ ∣ 0 ⟩ − cos 2 θ e i 2 ϕ ∣ 1 ⟩ ( 30 )
因此得到 r ⋅ σ ⃗ = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ \mathbf{r}\cdot\vec{\sigma} =\ket{r+}\bra{r+} - \ket{r-}\bra{r-} r ⋅ σ = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ 和 ∣ r + ⟩ ⟨ r + ∣ + ∣ r − ⟩ ⟨ r − ∣ = I \ket{r+}\bra{r+} + \ket{r-}\bra{r-} = \mathbb{I} ∣ r + ⟩ ⟨ r + ∣ + ∣ r − ⟩ ⟨ r − ∣ = I 。和 exercise2.2 相同,把 I \mathbb{I} I 和 r ⋅ σ ⃗ \mathbf{r}\cdot\vec{\sigma} r ⋅ σ 代入 ρ = ( I + r ⋅ σ ⃗ ) / 2 \rho=(\mathbb{I}+\mathbf{r}\cdot\vec{\sigma})/2 ρ = ( I + r ⋅ σ ) /2 得 ρ = ∣ r + ⟩ ⟨ r + ∣ \rho = \ket{r+}\bra{r+} ρ = ∣ r + ⟩ ⟨ r + ∣ 为纯态。
这道题我写的步骤更为详细,所以如果 exercise2.2 有看不懂的地方可以先看此题的解答,明白 r ⋅ σ ⃗ = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ \mathbf{r}\cdot\vec{\sigma} =\ket{r+}\bra{r+} - \ket{r-}\bra{r-} r ⋅ σ = ∣ r + ⟩ ⟨ r + ∣ − ∣ r − ⟩ ⟨ r − ∣ 是怎么回事就好了。
在本题中你会发现 r \mathbf{r} r 的定义不同于之前的 n ⃗ = ( x , y , z ) \vec{n} = (x,y,z) n = ( x , y , z ) 而是 r = ( sin θ cos ϕ , sin θ sin ϕ , cos θ ) \mathbf{r} = (\sin\theta\cos\phi,\sin\theta\sin\phi,\cos\theta) r = ( sin θ cos ϕ , sin θ sin ϕ , cos θ ) ,这是因为我们经常使用 Bloch 球来表示单个量子比特的几何图像:
在这样一个球坐标下的量子态表示为 Bloch 矢量 r \mathbf{r} r 。这个球的形状并不是我随意选用的,而是由量子态密度矩阵的半正定性 ∣ n ⃗ ∣ 2 = x 2 + y 2 + z 2 ≤ 1 \mid{\vec{n}}\mid^{2}=x^{2}+y^{2}+z^{2}\leq1 ∣ n ∣ 2 = x 2 + y 2 + z 2 ≤ 1 决定的,而借助 Bloch 球和半正定性我们可以对量子态进行直观的描述,比如在 Bloch 球球面上的所有矢量都是纯态(∣ r ∣ = 1 |\mathbf{r}| = 1 ∣ r ∣ = 1 ),之下皆为混合态(∣ r ∣ < 1 |\mathbf{r}| < 1 ∣ r ∣ < 1 ) 。
exercise 2.4
Q:试根据量子力学基本假设解释 Stern-Gerlach 实验(包括连续 S-G 实验)
A:由量子力学基本假设,测量自旋磁矩 Z 方向分量可用算符 S z = ℏ 2 σ z S_z = \frac\hbar2\sigma_z S z = 2 ℏ σ z 表示,同样 S x = ℏ 2 σ x S_x = \frac\hbar2\sigma_x S x = 2 ℏ σ x
实验测得只有两个斑点对应测量算符的两个本征值:
S z = ℏ 2 ∣ 0 ⟩ ⟨ 0 ∣ − ℏ 2 ∣ 1 ⟩ ⟨ 1 ∣ (31) S_z = \frac\hbar2\ket{0}\bra{0}-\frac\hbar2\ket{1}\bra{1}
\tag{31}
S z = 2 ℏ ∣ 0 ⟩ ⟨ 0 ∣ − 2 ℏ ∣ 1 ⟩ ⟨ 1 ∣ ( 31 )
只有两个取值,即两个实验结果,与实验相符。
连续测量自旋磁矩在 Z 方向分量
假定初始态为 ∣ ψ ⟩ = C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ \ket{\psi}=C_0\ket{0}+C_1\ket{1} ∣ ψ ⟩ = C 0 ∣ 0 ⟩ + C 1 ∣ 1 ⟩ ,则第一次测量 S z S_z S z 分别得到自旋向上、向下两种实验结果,其概率分别为:
↑ : ∣ ⟨ ψ ∣ 0 ⟩ ∣ 2 = ∣ C 0 ∣ 2 ↓ : ∣ ⟨ ψ ∣ 1 ⟩ ∣ 2 = ∣ C 1 ∣ 2 (32) \uparrow:|\lang\psi\ket{0}|^2 = |C_0|^2\quad\quad\downarrow:|\lang\psi\ket{1}|^2 = |C_1|^2
\tag{32}
↑: ∣ ⟨ ψ ∣ 0 ⟩ ∣ 2 = ∣ C 0 ∣ 2 ↓: ∣ ⟨ ψ ∣ 1 ⟩ ∣ 2 = ∣ C 1 ∣ 2 ( 32 )
然后只选取自旋向上的粒子,即初始态为 ∣ 0 ⟩ \ket{0} ∣ 0 ⟩ ,则再次测量 S z S_z S z 时,↑ : ∣ ⟨ 0 ∣ 0 ⟩ ∣ 2 = 1 , ↓ : ∣ ⟨ 0 ∣ 1 ⟩ ∣ 2 = 0 \uparrow:|\lang0\ket{0}|^2 = 1,\downarrow:|\lang0\ket{1}|^2 = 0 ↑: ∣ ⟨ 0 ∣ 0 ⟩ ∣ 2 = 1 , ↓: ∣ ⟨ 0 ∣ 1 ⟩ ∣ 2 = 0 ,只能观察到自旋向上的粒子,符合实验结果。
先测 S z S_z S z ,选取自旋向上粒子后再测 S x S_x S x
第一次测 S z S_z S z 的结果已验证,第二次测量的输入态确定为 ∣ 0 ⟩ \ket{0} ∣ 0 ⟩ ,则在测自旋向左右方向时得到两种结果:
S x = ∣ x + ⟩ ⟨ x + ∣ − ∣ x − ⟩ ⟨ x − ∣ (33) S_x = \ket{x+}\bra{x+}-\ket{x-}\bra{x-}
\tag{33}
S x = ∣ x + ⟩ ⟨ x + ∣ − ∣ x − ⟩ ⟨ x − ∣ ( 33 )
其概率分别为 ∣ x + ⟩ : ∣ ⟨ 0 ∣ x + ⟩ ∣ 2 = 1 2 , ∣ x − ⟩ : ∣ ⟨ 0 ∣ x − ⟩ ∣ 2 = 1 2 \ket{x+}:|\lang0\ket{x+}|^2 = \frac12,\ket{x-}:|\lang0\ket{x-}|^2 = \frac12 ∣ x + ⟩ : ∣ ⟨ 0 ∣ x + ⟩ ∣ 2 = 2 1 , ∣ x − ⟩ : ∣ ⟨ 0 ∣ x − ⟩ ∣ 2 = 2 1 ,符合实验结果。
在实验 3 的基础上选择 ∣ x + ⟩ \ket{x+} ∣ x + ⟩ 方向的粒子,再次测量 S z S_z S z 则得到实验结果的概率为 ↑ : ∣ ⟨ x + ∣ 0 ⟩ ∣ 2 = 1 2 , ↓ : ∣ ⟨ x + ∣ 1 ⟩ ∣ 2 = 1 2 \uparrow:|\lang x+\ket{0}|^2 = \frac12,\downarrow:|\lang x+\ket{1}|^2 = \frac12 ↑: ∣ ⟨ x + ∣ 0 ⟩ ∣ 2 = 2 1 , ↓: ∣ ⟨ x + ∣ 1 ⟩ ∣ 2 = 2 1 ,一共两种情况,和实验结果中出现两个斑点吻合。
说实话,看到这个解释我的第一反应是,扯淡 !!这不就是生凑出来的结果吗,根本没有什么严格的推导过程啊,不过后来想想,我们整个量子力学体系就是建构在五大公设上的,是假设出来的 ,我就释然了。。。
值得注意的是 σ x \sigma_x σ x 的本征分解结果:
∣ x + ⟩ = ( ∣ 0 ⟩ + ∣ 1 ⟩ ) / 2 ∣ x − ⟩ = ( ∣ 0 ⟩ − ∣ 1 ⟩ ) / 2 (34) \begin{align}
\ket{x+} &= (\ket{0}+\ket{1})/\sqrt{2}\\
\ket{x-} &= (\ket{0}-\ket{1})/\sqrt{2}
\end{align}
\tag{34}
∣ x + ⟩ ∣ x − ⟩ = ( ∣ 0 ⟩ + ∣ 1 ⟩ ) / 2 = ( ∣ 0 ⟩ − ∣ 1 ⟩ ) / 2 ( 34 )
顺便再把 σ y \sigma_y σ y 的本征分解结果放在这里:
∣ y + ⟩ = ( ∣ 0 ⟩ + i ∣ 1 ⟩ ) / 2 ∣ y − ⟩ = ( ∣ 0 ⟩ − i ∣ 1 ⟩ ) / 2 (35) \begin{align}
\ket{y+} &= (\ket{0}+i\ket{1})/\sqrt{2}\\
\ket{y-} &= (\ket{0}-i\ket{1})/\sqrt{2}
\end{align}
\tag{35}
∣ y + ⟩ ∣ y − ⟩ = ( ∣ 0 ⟩ + i ∣ 1 ⟩ ) / 2 = ( ∣ 0 ⟩ − i ∣ 1 ⟩ ) / 2 ( 35 )
exercise 2.5
Q:证明:
e i n ⋅ σ ⃗ ϕ = I cos ϕ + i ( n ⋅ σ ⃗ ) sin ϕ (36) e^{i\mathbf{n}\cdot\vec{\sigma}\phi} = \mathbb{I}\cos\phi+i(\mathbf{n}\cdot\vec{\sigma})\sin\phi
\tag{36}
e i n ⋅ σ ϕ = I cos ϕ + i ( n ⋅ σ ) sin ϕ ( 36 )
A:对于矩阵 M = n ⋅ σ ⃗ M = \mathbf{n}\cdot\vec{\sigma} M = n ⋅ σ 和复数 t = i ϕ t = i\phi t = i ϕ 有:
e M t = I + t M + t 2 2 ! M 2 + ⋯ = ∑ k = 0 ∞ 1 k ! t k M k (37) \begin{align}
e^{Mt}&=\mathbb{I}+tM+\frac{t^2}{2!}M^2+\cdots\\
&=\sum_{k = 0}^\infty\frac{1}{k!}t^kM^k
\end{align}
\tag{37}
e Mt = I + tM + 2 ! t 2 M 2 + ⋯ = k = 0 ∑ ∞ k ! 1 t k M k ( 37 )
利用 exercise 2.2 的结论,由于 ( n ⋅ σ ⃗ ) 2 = ∣ n ∣ 2 I (\mathbf{n}\cdot\vec{\sigma})^2 = |\mathbf{n}|^2\mathbb{I} ( n ⋅ σ ) 2 = ∣ n ∣ 2 I
e i n ⋅ σ ⃗ ϕ = I + i ϕ n ⋅ σ ⃗ − ϕ 2 2 ∣ n ∣ 2 I − i ϕ 3 6 n ⋅ σ ⃗ ∣ n ∣ 2 + ⋯ (38) e^{i\mathbf{n}\cdot\vec{\sigma}\phi} = \mathbb{I}+i\phi\mathbf{n}\cdot\vec{\sigma}-\frac{\phi^2}{2}|\mathbf{n}|^2\mathbb{I}-i\frac{\phi^3}{6}\mathbf{n}\cdot\vec{\sigma}|\mathbf{n}|^2+\cdots
\tag{38}
e i n ⋅ σ ϕ = I + i ϕ n ⋅ σ − 2 ϕ 2 ∣ n ∣ 2 I − i 6 ϕ 3 n ⋅ σ ∣ n ∣ 2 + ⋯ ( 38 )
不妨令 ∣ n ∣ = 1 |\mathbf{n}| = 1 ∣ n ∣ = 1 ,则:
e i n ⋅ σ ⃗ ϕ = I + i ϕ n ⋅ σ ⃗ − ϕ 2 2 I − i ϕ 3 6 n ⋅ σ ⃗ + ⋯ = I ( 1 − ϕ 2 2 ! + ϕ 4 4 ! + ⋯ ) + i ( n ⋅ σ ⃗ ) ( 1 − ϕ 3 3 ! + ϕ 5 5 ! + ⋯ ) = I cos ϕ + i ( n ⋅ σ ⃗ ) sin ϕ (39) \begin{align}
e^{i\mathbf{n}\cdot\vec{\sigma}\phi} &= \mathbb{I}+i\phi\mathbf{n}\cdot\vec{\sigma}-\frac{\phi^2}{2}\mathbb{I}-i\frac{\phi^3}{6}\mathbf{n}\cdot\vec{\sigma}+\cdots\\
&=\mathbb{I}(1-\frac{\phi^2}{2!}+\frac{\phi^4}{4!}+\cdots)+i(\mathbf{n}\cdot\vec{\sigma})(1-\frac{\phi^3}{3!}+\frac{\phi^5}{5!}+\cdots)\\
&=\mathbb{I}\cos\phi+i(\mathbf{n}\cdot\vec{\sigma})\sin\phi
\end{align}
\tag{39}
e i n ⋅ σ ϕ = I + i ϕ n ⋅ σ − 2 ϕ 2 I − i 6 ϕ 3 n ⋅ σ + ⋯ = I ( 1 − 2 ! ϕ 2 + 4 ! ϕ 4 + ⋯ ) + i ( n ⋅ σ ) ( 1 − 3 ! ϕ 3 + 5 ! ϕ 5 + ⋯ ) = I cos ϕ + i ( n ⋅ σ ) sin ϕ ( 39 )
证毕。
这个是幂有矩阵的欧拉公式,是个非常重要的公式,在之后的各种算符运算都离不开它。
这里说一下为什么可以令 ∣ n ∣ = 1 |\mathbf{n}| = 1 ∣ n ∣ = 1 。可以用于欧拉公式的矩阵需要满足一些条件:首先必须是方阵;其次,该矩阵必须可以用希尔伯特空间中的一组基进行线性展开,在二维希尔伯特空间中我们选用的基就是 Pauli 矩阵向量 σ ⃗ = ( σ x , σ y , σ z ) \vec{\sigma} = (\sigma_x,\sigma_y,\sigma_z) σ = ( σ x , σ y , σ z ) 。另外,就算是该矩阵用一组基展开之后的 ∣ n ∣ ≠ 1 |\mathbf{n}| \ne 1 ∣ n ∣ = 1 ,我们也可以从 ϕ \phi ϕ 里匀出来一个常数乘到 n \mathbf{n} n 里面使 ∣ n ∣ = 1 |\mathbf{n}| = 1 ∣ n ∣ = 1 。
感谢你能看完这篇文章,有误之处欢迎反馈!