ECE 4300: Lasers and Optoelectronics

Fall 2016, Debdeep Jena and Clif Pollock, Cornell University

Solutions to Assignment 5: By Kevin Lee, Athith Krishna and
Reet Chaudhuri

Problem: 5.1
Problem 5.2 (Verdeyen Problem # 8.32)
(a) and (b)
First we can calculate the loss in the system, 𝛼𝛼.
Without pumping (𝛾𝛾0 = 0), the transmission is 0.85.
𝑇𝑇 = 0.85 = 𝑒𝑒 −𝛼𝛼𝑙𝑙𝑔𝑔
−1
𝛼𝛼 = ln(0.85) = 3.25 × 10−4 𝑐𝑐𝑐𝑐−1
𝑙𝑙𝑔𝑔
For small signal gain, it means the optical intensity is much smaller than the
saturation intensity.
𝐼𝐼 ≪ 𝐼𝐼𝑠𝑠
In this case, we can use (8.3.18).
𝐼𝐼𝑜𝑜𝑜𝑜𝑜𝑜
𝐺𝐺0 = = 𝑒𝑒 𝑙𝑙𝑔𝑔(𝛾𝛾0 −𝛼𝛼) = 6.2807 = 7.98dB
𝐼𝐼𝑖𝑖𝑖𝑖
(c)
For a unity gain amplifier, it means that 𝐼𝐼𝑖𝑖𝑖𝑖 = 𝐼𝐼𝑜𝑜𝑜𝑜𝑜𝑜 . In this case, the derivative of the
intensity along z-axis is zero.
𝑑𝑑𝑑𝑑
=0
𝑑𝑑𝑑𝑑
𝛾𝛾0
− 𝛼𝛼 = 0
1 + 𝐼𝐼/𝐼𝐼𝑠𝑠
𝛾𝛾0 𝐼𝐼
= 1+
𝛼𝛼 𝐼𝐼𝑠𝑠
𝛾𝛾0
𝐼𝐼 = � − 1� 𝐼𝐼𝑠𝑠 = 180.923W/𝑐𝑐𝑐𝑐2
𝛼𝛼
ECE 4300: Lasers & Optoelectronics Name: Athith Krishna
Prof. Debdeep Jena Net-id: ak857

Solution:
(a) Given

𝐺0 = 6 𝑑𝐵 = 4 = exp ((𝛾0 − 𝛼)𝑙𝑔 )

Taking natural log on both sides,
1.386 1.386
(𝛾0 − 𝛼)𝑙𝑔 = 1.386 ⇒ 𝛾0 = −𝛼 = − 0.01 𝑐𝑚−1 = 0.1486 𝑐𝑚−1
𝑙𝑔 10𝑐𝑚

We know,
𝑔2 γ0 0.1486𝑐𝑚−1
𝛾0 = 𝜎 (𝑁2 − 𝑁1 ) ⇒ ΔN = = −20 𝑐𝑚 2
= 7.43 × 1018 𝑐𝑚−3
⏟ 𝑔1 𝜎 2 × 10
𝑖𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛=Δ𝑁

(b)
ℎ𝜈 ℎ𝑐 𝑘𝑊
𝐼𝑠 = = = 8.50
𝜎𝜏2 760 nm × 2 × 10−20 𝑐𝑚2 × 1.54 ms 𝑐𝑚2
(c) Using homogenous law,
1 𝑑𝐼 𝛾0 𝐼
( )( ) = − 𝛼 − −(1)
𝐼 𝑑𝑧 𝐼
1+𝐼
𝑠
𝐼
Let, 𝑥 = 𝐼 , then (1) can be re-written in terms 𝑥 as,
𝑠

𝛾0 𝛾
𝑑𝑥 [𝑥 − (( 𝛼 ) − 1)] [𝑥 − (( 𝛼0 ) − 1)]
= (−𝛼) = (−𝛼𝑥) − −(2)
𝑑𝑧 1 𝑥
To solve (2), we use method of separation of variables, then partial fractions, rewrite (2) as,

4
ECE 4300: Lasers & Optoelectronics Name: Athith Krishna
Prof. Debdeep Jena Net-id: ak857

1 1+𝑥
𝑑𝑧 = 𝑑𝑥 [ ] ⇒ 𝑑𝑧 =
𝛾 𝛾
[𝑥 − (( 𝛼0 ) − 1)] (−𝛼) [𝑥 2 − ((𝛼 ) − 1) 𝑥]

𝑋1 𝑋2
−𝛼 𝑑𝑧 = [ + ] 𝑑𝑥 − − − (3)
𝑥 𝑥 − ((𝛾0 ) − 1)
𝛼
1 𝛾0
Let 𝑋1 = − 𝛾 ; 𝑋2 = −𝑋1 ; and 𝜒 = − 1 = 1.386
( )−1 𝛼
𝛼

Integrating (3) with the limits,

𝑙𝑔 𝑥2 𝑥2
𝑑𝑥 𝑑𝑥
−(𝛾0 − 𝛼) ∫ 𝑑𝑧 = ∫ − ∫
𝑥−𝜒 𝑥
0 𝑥1 𝑥1

𝑥2 − 𝜒 𝑥2
−(𝛾0 − 𝛼)𝑙𝑔 = ln [ ] − ln [ ] − − − (4)
𝑥1 − 𝜒 𝑥1
𝐼𝑜𝑢𝑡
Given = 2, therefore,
𝐼𝑖𝑛

𝐼2 𝐼1
𝑥2 = 2𝑥1 = = 2( )
𝐼𝑠 𝐼𝑠
Taking exponential on both sides of (4),
2𝑥1 − 𝜒
exp[−(𝛾0 − 𝛼)𝑙𝑔 ] = 0.5 [ ] = 0.25 ⇒ 𝑥1 = 4.62
𝑥1 − 𝜒
𝐼1 𝑘𝑊
𝑥1 = ⇒ 𝐼1 = 39.2
𝐼𝑠 𝑐𝑚2

5
Problem 5.4 (Verdeyen Problem # 9.17)
(a)

Since there is no population in state 3, its lifetime is zero.

𝑁𝑁1 + 𝑁𝑁2 = 𝑁𝑁
To calculate the required power, we need to know the energy from state 1 to 3.
ℎ𝑐𝑐
= 4.417 × 10−19 𝐽𝐽
𝐸𝐸𝑔𝑔,31 =
𝜆𝜆
This is the energy required for pumping one electron from state 1 to 3.
𝑁𝑁2 = 1019 𝑐𝑐𝑐𝑐−3
For a steady state condition, the population of state 2 should be the product of
pumping rate and lifetime of state 2.
𝑁𝑁2 = 𝑅𝑅31 × 𝜏𝜏2
𝑅𝑅31 = 3.33 × 1021 𝑠𝑠 −1 𝑐𝑐𝑐𝑐−3
𝑉𝑉(𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣) = 1𝑐𝑐𝑐𝑐2 × 15𝑐𝑐𝑐𝑐 = 15𝑐𝑐𝑐𝑐3
Power = 𝑅𝑅31 × 𝐸𝐸𝑔𝑔,31 × 𝑉𝑉 = 22086𝑊𝑊
(b)
Spontaneous emission power can be easily calculated by following equation.
𝑁𝑁2
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 = × ℎ𝜈𝜈21 × 𝑉𝑉
𝜏𝜏2
ℎ𝜈𝜈21 : 𝑃𝑃ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 2 𝑡𝑡𝑡𝑡 1
Because this problem doesn’t specify the transition wavelength from state 2 to 1. I
am going to use the emission wavelength specified on page 293, which is 6943Å.
Put in all numbers into the equation.
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 = 14315.1𝑊𝑊
(c)
To calculate the maximum output power, we can utilize (9.4.13a).
ℎ𝜈𝜈𝑁𝑁𝑝𝑝 (max)
𝑃𝑃(max) = 𝜂𝜂𝑐𝑐𝑐𝑐𝑐𝑐
𝜏𝜏𝑝𝑝
First, let’s look for passive photon lifetime.
From (9.4.2c), we have the photon lifetime.
𝜏𝜏𝑅𝑅𝑅𝑅
𝜏𝜏𝑝𝑝 =
1 − 𝑆𝑆
2𝑙𝑙𝑔𝑔 2(𝑑𝑑 − 𝑙𝑙𝑔𝑔 )
𝜏𝜏𝑅𝑅𝑅𝑅 = 𝑐𝑐 + = 2.1𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛
�𝑛𝑛𝑔𝑔 𝑐𝑐
𝜏𝜏𝑅𝑅𝑅𝑅
𝜏𝜏𝑝𝑝 = = 6.3𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛
1 − 𝑅𝑅1 𝑅𝑅2
According to the 2nd equation in page 291, we can calculate the coupling efficiency.
𝑇𝑇2 0.3
𝜂𝜂𝑐𝑐𝑐𝑐𝑐𝑐 = = = 0.89522
1 − 𝑆𝑆 1 − 0.95 × 0.7
𝑛𝑛𝑖𝑖 − 𝑛𝑛𝑡𝑡ℎ 𝑛𝑛𝑡𝑡ℎ 𝑛𝑛𝑖𝑖
𝑁𝑁𝑝𝑝 (max) = − ln( )
2 2 𝑛𝑛𝑡𝑡ℎ
From (9.4.16a), we can calculate the initial atoms.
𝑛𝑛𝑖𝑖 = [𝑁𝑁2𝑖𝑖 − 𝑁𝑁1𝑖𝑖 ]𝐴𝐴𝑙𝑙𝑔𝑔 = [1019 − 0.58 × 1019 ] × 15 × 1 = 6.3 × 1019 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
To achieve threshold condition, we need to overcome the internal loss.
1 1
𝛾𝛾𝑡𝑡ℎ = ln � � = 𝜎𝜎(𝑁𝑁2 − 𝑁𝑁1 )𝑡𝑡ℎ
2𝑙𝑙𝑔𝑔 𝑅𝑅1 𝑅𝑅2
1 1
(𝑁𝑁2 − 𝑁𝑁1 )𝑡𝑡ℎ = ln � � = 1.07 × 1018 𝑐𝑐𝑐𝑐−3
2𝑙𝑙𝑔𝑔 𝜎𝜎 𝑅𝑅1 𝑅𝑅2
𝑛𝑛𝑡𝑡ℎ = (𝑁𝑁2 − 𝑁𝑁1 )𝑡𝑡ℎ 𝐴𝐴𝑙𝑙𝑔𝑔 = 1.6 × 1019 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
𝑁𝑁𝑝𝑝 (max) = 1.25 × 1019 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
𝑃𝑃(max) = 509.989𝑀𝑀𝑀𝑀
(d)
From (9.4.16c), we can calculate the output pulse energy.
𝑛𝑛𝑖𝑖 ℎ𝜈𝜈
𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜 = 𝜂𝜂𝑥𝑥𝑥𝑥𝑥𝑥 𝜂𝜂𝑐𝑐𝑐𝑐𝑐𝑐
2
Extraction efficiency can only be solved analytically from (9.4.14).
𝑛𝑛𝑖𝑖 6.3 × 1019
= = 3.9375
𝑛𝑛𝑡𝑡ℎ 1.6 × 1019
From FIGURE 9.13, extraction efficiency is around 0.98.
𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜 = 7.9121 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽
(e)
𝑊𝑊𝑜𝑜𝑜𝑜𝑜𝑜 7.9121𝐽𝐽
∆𝑡𝑡 ≈ = = 15.51𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛
𝑃𝑃0 (max) 509.989𝑀𝑀𝑀𝑀
From FIGURE 9.14, the approximate linewidth is 2.5.
∆𝑡𝑡 = 2.5 × 𝜏𝜏𝑝𝑝 = 15.75𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛
(f)
ECE 4300: Lasers & Optoelectronics Name: Athith Krishna
Prof. Debdeep Jena Net-id: ak857

Solution:
(a)
2
𝜏𝑅𝑇 [ 𝑐 ∙ (𝑛1 (15𝑐𝑚) + 𝑛2 (5𝑐𝑚) + 𝑛3 (10𝑐𝑚))]
𝜏𝑝 = =
1−𝑆 [1 − 𝑅1 𝑅2 𝑇 8 ]
2
[ 𝑐 ∙ ({1}(15𝑐𝑚) + {2.3}(5𝑐𝑚) + {1.816}(10𝑐𝑚))]
= = 9.97 × 10−9 𝑠
[1 − (0.95 ∙ 0.8 ∙ (0.99)8 )]
= 9.97 𝑛𝑠
At threshold,

9
ECE 4300: Lasers & Optoelectronics Name: Athith Krishna
Prof. Debdeep Jena Net-id: ak857
1 1 1 1 1 1
𝛾𝑡ℎ = ln = ln 8
= ln = 1.78 × 10−2 𝑐𝑚−1
2lg 𝑆 2lg 𝑅1 𝑅2 𝑇 2(10) 0.95 ∙ 0.8 ∙ (0.99)8

Using values from the table,

𝐵𝑟𝑎𝑛𝑐ℎ𝑖𝑛𝑔 𝑅𝑎𝑡𝑖𝑜 0.0799
𝐴21 = 4 = = 313.3 𝑠 −1
𝐹𝑙𝑢𝑜𝑟𝑜𝑠𝑐𝑒𝑛𝑡 𝑙𝑖𝑓𝑒𝑡𝑖𝑚𝑒 𝑜𝑓 𝐹3 255𝜇𝑠
2

𝐴21 𝜆2 2 313.3 𝑠 −1 (𝜆2 ) 2

𝜎= 2
( ) = 2
( ) = 2.51 × 10−19 𝑐𝑚2
8𝜋𝑛 𝜋Δ𝜈 8𝜋𝑛 𝜋(108 × 109 Hz)
It is given that the transition between the levels is from 11,423 𝑐𝑚−1 to 2002 𝑐𝑚−1,
𝑛
It is also given that, 𝑛 𝑖 = 3
𝑡ℎ

Initially,
𝑁2 𝐸2 − 𝐸1
= exp (− )
𝑁1 𝑘𝑇
𝑁3 𝐸3 − 𝐸1
= exp (− )
𝑁1 𝑘𝑇
[𝑁]𝑙𝑜𝑤𝑒𝑟 = 7.09 × 1016 𝑐𝑚−3
[𝑁]𝑢𝑝𝑝𝑒𝑟 = 4.73 × 1016 𝑐𝑚−3

Using Eq. 9.4.14,

[𝑁]𝑡ℎ = 3.53 × 1017 𝑐𝑚−3
From this, we get 𝑛𝑡ℎ , 𝑛𝑖
𝑛𝑡ℎ = 7.67 × 1017 #
𝑛𝑖 = 3𝑛𝑡ℎ = 2.3 × 1018 #
𝑛𝑖 𝑛𝑖 −𝑛𝑓
As we know = 3, we can use Fig 9.13 in Verdeyen to get the 𝜂𝑥𝑡𝑛 = = 0.941
𝑛𝑡ℎ 𝑛𝑖

𝑇2 0.2 0.2
𝜂𝑐𝑝𝑙 = = = = 0.669
1 − 𝑆 1 − 𝑅1 𝑅2 𝑇 8 1 − 0.95 ∙ 0.8 ∙ (0.99)8
[𝑁]𝑓𝑖𝑛𝑎𝑙 = 2.1 × 1016 𝑐𝑚−3

(b)
From part (a), we know 𝜂𝑐𝑝𝑙 and 𝜂𝑥𝑡𝑛 ,

𝑛𝑖 ℎ𝜈 2.3 × 1018 ℎ𝜈
𝑊 = 𝜂𝑥𝑡𝑛 𝜂𝑐𝑝𝑙 ( ) = (0.941)(0.669) ( ) = 0.1353 𝐽
2 2

10
ECE 4300: Lasers & Optoelectronics Name: Athith Krishna
Prof. Debdeep Jena Net-id: ak857
Using Eq. 9.4.13a
ℎ𝜈 ℎ𝜈 𝑛𝑖 − 𝑛𝑡ℎ 𝑛𝑡ℎ 𝑛𝑖
𝑃𝑚𝑎𝑥 = 𝜂𝑐𝑝𝑙 ∙ (𝑁𝑝 𝑚𝑎𝑥 ) = 𝜂𝑐𝑝𝑙 ∙ [ − ( ) ln ( )]
𝜏𝑝 𝜏𝑝 2 2 𝑛𝑡ℎ
18 17
ℎ𝜈 2.3 × 10 − 7.67 × 10 7.67 × 1017
= 0.669 ∙ [ −( ) ln(3)]
𝜏𝑝 2 2
= 4.33 × 106 𝑊
𝑛
From Fig 9.14, and 𝑛 𝑖 = 3 , we can calculate the pulse width
𝑡ℎ

Δ𝑡1
2
𝐹𝑊𝐻𝑀 = = 5.5 − 2.3 = 3.2 ⇒ Δ𝑡1 = (3.2)𝜏𝑝 = 31.9 𝑛𝑠
𝜏𝑝 2

We can compare this to the result obtained from Eq. 9.4.17,

𝑊 0.1353 𝐽
Δ𝑡1 = = ≈ 31.2 𝑛𝑠
2 𝑃𝑚𝑎𝑥 4.33 × 106 𝑊

Solution:
Given, 𝜆0 = 588.9 𝑛𝑚 and 𝜎 = 10−14
Using eq. 9.6.3b,
ℎ𝜈 ℎ𝑐 𝐽
𝑤𝑠 = = = 16.86 × 10−6
2𝜎 2𝜆𝜎 𝑐𝑚2
𝐺0 = −30𝑑𝐵 = 10−3
Using Eq 9.6. _ (on P.315),

11
Problem 5.6 (Verdeyen Problem # 9.23)
To calculate the output intensity, we should use (9.6.11).
𝐺𝐺0 𝑒𝑒 𝑢𝑢1 (𝑡𝑡)
𝐼𝐼2 (𝑡𝑡) = 𝐼𝐼1 (𝑡𝑡)
1 + 𝐺𝐺0 [𝑒𝑒 𝑢𝑢1 (𝑡𝑡) − 1]
The small signal gain is already specified -30dB.
𝐺𝐺0 = 10−3
Now we have to look for the pulse energy normalized by the characteristic saturation
energy.
𝓌𝓌(𝑡𝑡)
𝑢𝑢1 (𝑡𝑡) =
𝓌𝓌𝑠𝑠
𝑡𝑡

𝓌𝓌(𝑡𝑡) = � 𝐼𝐼1 (𝑡𝑡)𝑑𝑑𝑑𝑑

−∞

ℎ𝜈𝜈
𝓌𝓌𝑠𝑠 = = 3.375 × 10−5 𝐽𝐽/𝑐𝑐𝑐𝑐2 = 33.75𝜇𝜇𝜇𝜇/𝑐𝑐𝑐𝑐2
𝜎𝜎
𝓌𝓌 𝜋𝜋𝜋𝜋 𝓌𝓌 2𝜋𝜋𝜋𝜋
𝐼𝐼1 (𝑡𝑡) = 𝑠𝑠𝑠𝑠𝑠𝑠2 � � = [1 − cos( )]
𝑇𝑇 𝑇𝑇 2𝑇𝑇 𝑇𝑇
𝑡𝑡
𝓌𝓌 𝑇𝑇 2𝜋𝜋𝜋𝜋 𝓌𝓌 𝑡𝑡 1 2𝜋𝜋𝑡𝑡
𝓌𝓌(𝑡𝑡) = � 𝐼𝐼1 (𝑡𝑡)𝑑𝑑𝑑𝑑 = �𝑡𝑡 − sin � �� = [ − sin � �]
2𝑇𝑇 2𝜋𝜋 𝑇𝑇 2 𝑇𝑇 2𝜋𝜋 𝑇𝑇
−∞

Plug in all the functions into the equation of output intensity.

Problem 5.7 (Verdeyen Problem # 9.27)
(a)
The power spectrum is given in the problem.
𝑛𝑛=∞
𝑛𝑛𝜔𝜔𝑐𝑐
𝑃𝑃(𝜔𝜔) = 𝑃𝑃0 � 𝑠𝑠𝑠𝑠𝑠𝑠ℎ2 � � 𝛿𝛿(𝜔𝜔 − [𝜔𝜔0 + 𝑛𝑛𝜔𝜔𝑐𝑐 ])
∆𝜔𝜔
𝑛𝑛=−∞

To calculate the average power, we need to find instantaneous power. And it can be
related to electric field.
𝑒𝑒(𝜔𝜔) × 𝑒𝑒(𝜔𝜔)∗
𝑃𝑃(𝜔𝜔) = 𝐼𝐼(𝜔𝜔) × 𝐴𝐴 = 𝐴𝐴
2𝜂𝜂0
And we can do an inverse Fourier transform to find the time-domain function of the
electric field.
+∞
2𝜂𝜂0 𝑃𝑃0 𝑗𝑗𝜔𝜔 𝑡𝑡 𝑛𝑛𝜔𝜔𝑐𝑐 𝑗𝑗𝑛𝑛𝜔𝜔 𝑡𝑡
𝑒𝑒(𝑡𝑡) = � 𝑒𝑒 0 � sech( )𝑒𝑒 𝑐𝑐
𝐴𝐴 ∆𝜔𝜔
𝑛𝑛=−∞

Now I have to evaluate that big summation. As shown in p.301, we can approximate
this summation with integral.
Let
𝑛𝑛𝑛𝑛𝑐𝑐 = 𝑥𝑥
𝑑𝑑𝑑𝑑� = 1
𝜔𝜔𝑐𝑐

+∞
𝑛𝑛𝜔𝜔𝑐𝑐 𝑗𝑗𝑛𝑛𝑛𝑛 𝑡𝑡 1 ∞ 𝑥𝑥
� sech( )𝑒𝑒 𝑐𝑐 ~ � sech( )𝑒𝑒 𝑗𝑗𝑗𝑗𝑗𝑗 𝑑𝑑𝑑𝑑
∆𝜔𝜔 𝜔𝜔𝑐𝑐 −∞ ∆𝜔𝜔
𝑛𝑛=−∞

Since the sech function is an even function, we can further simplify the integral.
1 ∞ 𝑥𝑥 𝑗𝑗𝑗𝑗𝑗𝑗 1 ∞ 𝑥𝑥
� sech( )𝑒𝑒 𝑑𝑑𝑑𝑑 = � sech � � �cos(𝑥𝑥𝑥𝑥) + 𝑗𝑗𝑗𝑗𝑗𝑗𝑗𝑗(𝑥𝑥𝑥𝑥)�𝑑𝑑𝑑𝑑
𝜔𝜔𝑐𝑐 −∞ ∆𝜔𝜔 𝜔𝜔𝑐𝑐 −∞ ∆𝜔𝜔
2 ∞ 𝑥𝑥
= � sech � � cos(𝑥𝑥𝑥𝑥)𝑑𝑑𝑑𝑑
𝜔𝜔𝑐𝑐 0 ∆𝜔𝜔
Solve this integral by Mathematica.
2 ∞ 𝑥𝑥 2 1 𝜋𝜋∆𝜔𝜔𝜔𝜔 𝜋𝜋∆𝜔𝜔 𝜋𝜋∆𝜔𝜔𝜔𝜔
� sech( )𝑒𝑒 𝑗𝑗𝑗𝑗𝑗𝑗 𝑑𝑑𝑑𝑑 = × 𝜋𝜋∆𝜔𝜔 × sech � �= sech � �
𝜔𝜔𝑐𝑐 −∞ ∆𝜔𝜔 𝜔𝜔𝑐𝑐 2 2 𝜔𝜔𝑐𝑐 2

2𝜂𝜂0 𝑃𝑃0 𝑗𝑗𝜔𝜔 𝑡𝑡 𝜋𝜋∆𝜔𝜔 𝜋𝜋∆𝜔𝜔𝜔𝜔

𝑒𝑒(𝑡𝑡) = � 𝑒𝑒 0 sech � �
𝐴𝐴 𝜔𝜔𝑐𝑐 2

𝜋𝜋∆𝜔𝜔 2 𝜋𝜋∆𝜔𝜔𝜔𝜔
𝑝𝑝(𝑡𝑡) = 𝑃𝑃0 � � 𝑠𝑠𝑠𝑠𝑠𝑠ℎ2 ( )
𝜔𝜔𝑐𝑐 2
Finally, we can calculate the average power.
 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑤𝑤𝑤𝑤𝑤𝑤ℎ𝑖𝑖𝑖𝑖 𝑜𝑜𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 =
𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
The period can be easily calculated through rough trip time.
2𝜋𝜋
𝜏𝜏𝑅𝑅𝑅𝑅 =
𝜔𝜔𝑐𝑐
∞
∫ 𝑝𝑝(𝑡𝑡)𝑑𝑑𝑑𝑑
〈𝑃𝑃〉 = −∞ = 200𝑚𝑚𝑚𝑚
𝜏𝜏𝑅𝑅𝑅𝑅
(b)
From the above equation, we have the peak power when sech equals to 1.
𝜋𝜋∆𝜔𝜔 2
𝑃𝑃𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 = 𝑃𝑃0 � � = 9869.6𝑊𝑊
𝜔𝜔𝑐𝑐
(c)
It’s obvious that the lineshape is determined by the sech function. So FWHM is just
the sech function equals to half.
𝜋𝜋∆𝜔𝜔𝜔𝜔 1
𝑠𝑠𝑠𝑠𝑠𝑠ℎ2 � �=
2 2
𝜋𝜋∆𝜔𝜔𝜔𝜔 1
sech � �=
2 √2
Solve it by numerical method. We get
𝑡𝑡1 = 8.93 × 10−11 𝑠𝑠
2

𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 = 2𝑡𝑡1 = 0.1786𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛

2
Problem 5.8 (Verdeyen Problem # 9.31)
(a)
This problem is similar to Verdeyen Problem # 9.6c.
𝑑𝑑𝑑𝑑 𝛾𝛾0
={ − 𝛼𝛼}𝑓𝑓
𝑑𝑑𝑑𝑑 1 + 𝑓𝑓
𝑑𝑑𝑑𝑑 𝛾𝛾0 − 𝛼𝛼(1 + 𝑓𝑓)
= 𝑓𝑓
𝑑𝑑𝑑𝑑 1 + 𝑓𝑓
1 + 𝑓𝑓
𝑑𝑑𝑑𝑑 = 𝑑𝑑𝑑𝑑
𝑓𝑓(𝛾𝛾0 − 𝛼𝛼(1 + 𝑓𝑓))
1 + 𝑓𝑓 𝑐𝑐 𝑑𝑑
= +
𝛾𝛾0 − 𝛼𝛼(1 + 𝑓𝑓) 𝑓𝑓 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝛼𝛼
1
𝑐𝑐 =
𝛾𝛾0 − 𝛼𝛼
𝛾𝛾0
𝑑𝑑 =
(𝛾𝛾0 − 𝛼𝛼)
1 𝛾𝛾0
� + � 𝑑𝑑𝑑𝑑 = (𝛾𝛾0 − 𝛼𝛼)𝑑𝑑𝑑𝑑
𝑓𝑓 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝛼𝛼
𝑓𝑓2
1 𝛾𝛾0
� � + � 𝑑𝑑𝑑𝑑 = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝑓𝑓 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝛼𝛼
𝑓𝑓2 𝛾𝛾0 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝑓𝑓2
ln � � − ln � � = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝛼𝛼 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝑓𝑓1
So the problem is asking for extractable power, meaning the additional power we get
from the amplifier. Let’s assume 𝑓𝑓2 = 𝑓𝑓1 + 𝑓𝑓𝑒𝑒 .
The above equation should be rewritten.
𝑓𝑓𝑒𝑒 𝛾𝛾0 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝑓𝑓2
ln �1 + � − ln � � = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝛼𝛼 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝑓𝑓1
And we use an approximation for ln(1 + 𝑥𝑥)~𝑥𝑥. Since we’re considering the
maximum extractable power, it’s reasonable to assume the 𝑓𝑓1 is a large number,
meaning the approximation is valid.
𝑓𝑓𝑒𝑒 𝛾𝛾0 𝛼𝛼 𝛾𝛾0 𝛼𝛼
− ln �1 − 𝑓𝑓2 � + ln �1 − 𝑓𝑓 � = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 1
𝑓𝑓𝑒𝑒 𝛾𝛾0 𝛼𝛼 𝛾𝛾0 𝛼𝛼 𝑓𝑓𝑒𝑒 𝛾𝛾0 𝛼𝛼 𝛾𝛾0 𝛼𝛼
+ 𝑓𝑓2 − 𝑓𝑓1 = + (𝑓𝑓1 + 𝑓𝑓𝑒𝑒 ) − 𝑓𝑓
𝑓𝑓1 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 𝑓𝑓1 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 1
𝑓𝑓𝑒𝑒 𝛾𝛾0 𝛼𝛼
= + 𝑓𝑓 = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 𝑒𝑒
𝑓𝑓𝑒𝑒 𝛾𝛾0 𝛼𝛼 𝛾𝛾0
+ 𝑓𝑓𝑒𝑒 ~ 𝑓𝑓 = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 𝛾𝛾0 − 𝛼𝛼 𝑒𝑒
 (𝛾𝛾0 − 𝛼𝛼)2
𝑓𝑓𝑒𝑒 = 𝑙𝑙𝑔𝑔
𝛾𝛾0
(𝛾𝛾0 − 𝛼𝛼)2
𝐼𝐼𝑒𝑒,𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑙𝑙𝑔𝑔 × 𝐼𝐼𝑠𝑠
𝛾𝛾0
(b)
Now we have to apply it into a ring laser.
First, we defined a few parameters according to the problem.
𝑔𝑔𝑡𝑡ℎ = (𝛾𝛾𝑡𝑡ℎ − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑔𝑔0 = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝛼𝛼
𝛿𝛿 =
𝛾𝛾0 − 𝛼𝛼
From (9.2.4), input and output intensities can be related.
𝑓𝑓1 = 𝑆𝑆𝑓𝑓2
𝑆𝑆 = 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓
In order to reach threshold condition, we have one more relation.
𝑆𝑆 × 𝑒𝑒 𝑔𝑔𝑡𝑡ℎ = 1
1
𝑔𝑔𝑡𝑡ℎ = ln � � = −ln(𝑆𝑆)
𝑆𝑆
𝑓𝑓1 = 𝑆𝑆𝑓𝑓2 = 𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ 𝑓𝑓2

Plug in all the parameters into following equation we solved in the previous problem.
𝑓𝑓2 𝛾𝛾0 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝑓𝑓2
ln � � − ln � � = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝛼𝛼 (𝛾𝛾0 − 𝛼𝛼) − 𝛼𝛼𝑓𝑓1
𝑓𝑓2 𝛿𝛿 + 1 1 − 𝛿𝛿𝑓𝑓1
ln � � + ln( ) = (𝛾𝛾0 − 𝛼𝛼)𝑙𝑙𝑔𝑔
𝑓𝑓1 𝛿𝛿 1 − 𝛿𝛿𝛿𝛿2
𝛿𝛿 + 1 1 − 𝛿𝛿𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ 𝑓𝑓2
𝑔𝑔𝑡𝑡ℎ + ln( ) = 𝑔𝑔0
𝛿𝛿 1 − 𝛿𝛿𝛿𝛿2
1 − 𝛿𝛿𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ 𝑓𝑓2 𝛿𝛿
ln � �= (𝑔𝑔 − 𝑔𝑔𝑡𝑡ℎ )
1 − 𝛿𝛿𝛿𝛿2 𝛿𝛿 + 1 0
1 − 𝛿𝛿𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ 𝑓𝑓2 𝛿𝛿
= 𝑒𝑒 𝛿𝛿+1(𝑔𝑔0 −𝑔𝑔𝑡𝑡ℎ )
1 − 𝛿𝛿𝛿𝛿2
𝛿𝛿
1 − 𝑒𝑒 𝛿𝛿+1(𝑔𝑔0 −𝑔𝑔𝑡𝑡ℎ )
𝑓𝑓2 = 𝛿𝛿
𝛿𝛿(𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ − 𝑒𝑒 𝛿𝛿+1(𝑔𝑔0 −𝑔𝑔𝑡𝑡ℎ ) )
Output can be calculated from (9.2.6).
𝛿𝛿
1 − 𝑒𝑒 𝛿𝛿+1(𝑔𝑔0 −𝑔𝑔𝑡𝑡ℎ )
𝐼𝐼𝑜𝑜𝑜𝑜𝑜𝑜 = 𝑇𝑇𝑏𝑏 𝑇𝑇2 𝐼𝐼𝑠𝑠 𝑓𝑓2 = 𝑇𝑇𝑏𝑏 𝑇𝑇2 𝐼𝐼𝑠𝑠 𝛿𝛿
𝛿𝛿(𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ − 𝑒𝑒 𝛿𝛿+1(𝑔𝑔0 −𝑔𝑔𝑡𝑡ℎ ) )
Now if 𝛿𝛿 ≪ 1, it means the factor power of exp is very small. We can make the
following approximation.
𝑒𝑒 𝑥𝑥 ≈ 1 + 𝑥𝑥 𝑖𝑖𝑖𝑖 𝑥𝑥 ≪ 1
𝛿𝛿
1−1− (𝑔𝑔 − 𝑔𝑔𝑡𝑡ℎ ) 1 (𝑔𝑔0 − 𝑔𝑔𝑡𝑡ℎ )
𝐼𝐼𝑜𝑜𝑜𝑜𝑜𝑜 = 𝑇𝑇𝑏𝑏 𝑇𝑇2 𝐼𝐼𝑠𝑠 𝛿𝛿 + 1 0 ≈ 𝑇𝑇𝑏𝑏 𝑇𝑇2 𝐼𝐼𝑠𝑠
𝛿𝛿 𝛿𝛿 + 1 (1 − 𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ )
𝛿𝛿(𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ − 1 − (𝑔𝑔0 − 𝑔𝑔𝑡𝑡ℎ ))
𝛿𝛿 + 1
𝛾𝛾0 − 𝛼𝛼 (𝑔𝑔0 − 𝑔𝑔𝑡𝑡ℎ )
= 𝑇𝑇𝑏𝑏 𝑇𝑇2 𝐼𝐼𝑠𝑠
𝛾𝛾0 (1 − 𝑒𝑒 −𝑔𝑔𝑡𝑡ℎ )
Problem 5.9 (Verdeyen Problem # 9.36)
(a)
𝑐𝑐
= 200𝑀𝑀𝑀𝑀𝑀𝑀
2𝑑𝑑
𝑐𝑐
𝑑𝑑 = = 0.75𝑚𝑚
2 × 200𝑀𝑀𝑀𝑀𝑀𝑀
(b)
2𝑑𝑑
𝑇𝑇 = = 5𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛
𝑐𝑐
(c)
𝐸𝐸(𝑡𝑡) × 𝐸𝐸(𝑡𝑡)∗
𝐼𝐼(𝑡𝑡) =
2𝜂𝜂0
Let’s first try to find the analytical form of electric field.
∞ 𝑛𝑛𝜔𝜔
sin[ ∆𝜔𝜔𝑐𝑐 ]𝑒𝑒 𝑗𝑗𝑗𝑗𝜔𝜔𝑐𝑐𝑡𝑡
𝐸𝐸(𝑡𝑡) = 𝐸𝐸0 𝑒𝑒 𝑗𝑗𝜔𝜔0 𝑡𝑡 � 𝑛𝑛𝜔𝜔𝑐𝑐
−∞ ∆𝜔𝜔
Now we’re using the same trick to replace the summation with integral.
∞ 𝑛𝑛𝜔𝜔 𝑥𝑥
sin[ ∆𝜔𝜔𝑐𝑐 ]𝑒𝑒 𝑗𝑗𝑗𝑗𝜔𝜔𝑐𝑐𝑡𝑡 ∞ sin[ ] 𝑑𝑑𝑑𝑑
� = � ∆𝜔𝜔 𝑒𝑒 𝑗𝑗𝑗𝑗𝑗𝑗
𝑛𝑛𝜔𝜔𝑐𝑐 𝑥𝑥 𝜔𝜔𝑐𝑐
−∞
−∞ ∆𝜔𝜔 ∆𝜔𝜔
By observation, the original time domain should be a square pulse function, because
the Fourier transform of square pulse function is sinc function.
𝑥𝑥 𝑥𝑥
−1
∞ sin[
∆𝜔𝜔
]
𝑗𝑗𝑗𝑗𝑗𝑗 −1
∆𝜔𝜔 1 ∞ 2 sin �∆𝜔𝜔� 𝑗𝑗𝑗𝑗𝑗𝑗
𝜔𝜔𝑐𝑐 � 𝑥𝑥 𝑒𝑒 𝑑𝑑𝑑𝑑 = 2𝜋𝜋𝜔𝜔𝑐𝑐 [ � 𝑥𝑥 𝑒𝑒 𝑑𝑑𝑑𝑑]
−∞ 2 2𝜋𝜋 −∞ ∆𝜔𝜔
∆𝜔𝜔 ∆𝜔𝜔
∆𝜔𝜔 ∆𝜔𝜔
= 2𝜋𝜋𝜔𝜔𝑐𝑐 −1 × 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟( 𝑡𝑡)
2 2
∆𝜔𝜔 1
𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 � 𝑡𝑡� = 1 𝑓𝑓𝑓𝑓𝑓𝑓|𝑡𝑡| <
2 ∆𝜔𝜔
∆𝜔𝜔 1
𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 � 𝑡𝑡� = 0 𝑓𝑓𝑓𝑓𝑓𝑓|𝑡𝑡| >
2 ∆𝜔𝜔
𝐸𝐸0 2 2 ∆𝜔𝜔 2 ∆𝜔𝜔 2
𝐼𝐼(𝑡𝑡) = 𝜋𝜋 ( ) (𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟( 𝑡𝑡))
2𝜂𝜂0 𝜔𝜔𝑐𝑐 2
Now plot it with Mathematica.
This pulse should repeat for a period of 5nsec.
𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 = 81868𝑊𝑊/𝑐𝑐𝑐𝑐2
(c)
This is a simple question. The output pulse energy should be the product of the
transmission function and the pulse itself.
1
∆𝜔𝜔 𝜋𝜋𝜋𝜋
𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 1 = 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 � 𝑐𝑐𝑐𝑐𝑐𝑐 2 � � 𝑑𝑑𝑑𝑑
−1 𝑇𝑇
∆𝜔𝜔

2
If there is no shutter, meaning 𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑡𝑡 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 0 = 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 ∆𝜔𝜔

𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 0 − 𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 1

𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 = = 0.000133
𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 0