MIT 2.111/8.411/6.898/18.

435 Quantum Information Science I Sam Ocko

Fall, 2010 September 29, 2010

Problem Set # 3 Solutions

1. Measurements and Uncertainty (a) The expectation value of M on state | will be m, the standard deviation will be 0. (b) Measuring X on the state |0 , we will get results of 1,-1 with equal probability. Therefore the expectation value is 0 and the standard deviation is 1. 2. Entropy of quantum states (a) The entropy of 0 = |0 0| is log2 (1) = 0 (b) The entropy of 1 = (c) If T r(2 ) = 1, 2 k =
k k (|0 0|+|1 1|) 2 1 1 1 is 1 2 log2 ( 2 ) 2 log2 ( 2 ) = 1

k = 1

Therefore, k (k 1) = 0
k

Since 0 k 1, k , we know that k (k 1) 0, k , and thus the only way for the above condition to be satised is for k = 0, 1, k . Therefore T r(2 ) = 1 if and only if has a single eigenvalue of 1 with all other eigenvalues 0.

S ( ) =
k

k log2 (k ) = 0

Since 0 k 1, k , we know that k log2 (k ) 0, k . Therefore, the only way for the above condition to be satised is for k = 0, 1, k , and thus S () = 1 if and only if has a single eigenvalue of 1 with all other eigenvalues 0. Therefore, for density matrices, T r(2 ) = 1 and S () = 1 are equivalent statements. 3. (a) A state is a product state if and only if it can be represented as | = |A |B . If a state has a Schmidt number 1, it can be represented as a product state k |kA |kB = |A |B since only one Schmidt coecient is nonzero. If it has k a Schmidt number greater than 1, it has no such representation as |A |B , because if it did it would have a Schmidt number of 1 through the above representation.

2 (b) Lemma: If an entangled state between Alice and Bob has the Schmidt decomposition k |kA |kB
k

Then Alices reduced density matrix is A =

k

k |kA kA |

(Likewise for Bob) Therefore, if | has a Schmidt number of 1, the reduced density matrices A , B have only one non-zero eigenvalue and are pure states. If | has a Schmidt number greater than 1, the reduced density matrices A , B have multiple non-zero eigenvalues and are mixed states. Proof Of Lemma(Less Mathematical) If Bob measured his state in the {|kB } basis, with probability k he will measure |kB and Alices state will collapse to |kA . Therefore, since the outcome of Alices measurement cant be eected by whether Bob made his measurement, we may say: A =
k

k |kA kA |

Proof of Lemma(More Mathematical): We may write the global density matrix as total =
k,K

k k |kA

|kB

kB |B kA |A

Alices reduced density matrix can be obtained by taking the partial trace A = T rB (total ) =
k,K

k k |kA kA | T r (|kB

kB |)

Since T r (|kB

kB |) = kB | |kB = kk , we may say A =

k

k |kA kA |
1 3

11 +|22 4. (a) The Schmidt decomposition of |1 = |00 +| is 3 tion(Other Schmidt decompositions are also possible)

k=0,1,2

|k |k by inspec-

(b) The Schmidt decomposition of |2 = 1 |+ = (|0 + |1 ) 2

|00 +|01 +|10 +|11 2

is |+ |+ by inspection, where

|10 |11 1 (c) The Schmidt decomposition of |3 = |00 +|01 + is (|0 |+ + |1 | ) by in2 2 1 spection, where | = 2 (|0 |1 ) (Other Schmidt decompositions are also possible) 01 +|11 (d) To nd the schmidt decomposition of |4 = |00 +| , we use the lemma proved in 3 the previous problem. We can see that Alices reduced density matrix is

A =

1 3

2 1 1 1

3 having eigenvalues 1 +
1 2 1 6

3+

1 5 .87, 6 3

5 .13, eigenvectors 1 +
1 2

3+
1 2

1 + 1 +

5 ,1 3+ 5

(.85, .52),

3
1 2

1 + 1 +

5 ,1 3 5

(.52, .85)

Likewise, Bobs reduced density matrix is 1 3 1 1 1 2 5 .13, eigenvectors 1, 1

1 2 1 2

B = having eigenvalues 1, 1 +
1 2 1 2 1 6

3+

1 5 .87, 6 3

3+

1 + 1 +

3+

(.52, .85),

1 + 1 +

(.85, .52)

Therefore, the Schmidt decomposition will be |00 + |01 + |11 = 1 |A1 |B 1 + 2 |A2 |B 2 3 5 .87, 2 = 1 5 .13 Where 1 = 1 6 3+ 6 3 And the Schmidt vectors |A1 , |A2 , |B 1 , |B 2 are dened as above |4 = 5. (a) 1 = 2 = 3 =
1 2 1 2 1 2

(|0 + |1 ) |0 + ei |1 1 + ei |0 + (1 ei |1
1cos() 2

i 2 (b) The probability of measuring a qubit to be 1 is p = 1 4 |1 e | =

(c) We consider the random variable X which is 0 if we measure 0 and 1 if we measure 1. () X = X 2 = 1cos Therefore, the variance of a single measurement is 2 X2 X
2

1 cos() 2

1 cos() 2

sin()2 4
2

) Thus the variance in the number of 1s you get after n measurements is n sin( 4 and thus the standard deviation of measured probability after n experiments will be | sin()| 2 n

Since

1 = n | (Note that its impossible to tell the sign of in this way, you need to measure in a dierent basis to do that)

dp d

sin() 2 ,

the accuracy of the estimate for is =

p dp/d

| sin()| 2 n sin() 2