Density Operator of a Two-Level System Electric Dipole Interaction

Laser Physics - Exercise class # 1 September 7, 2012

Pure Case

Let us consider a two-level system. The lower level is called |a and the upper level is called |b . We call 0 the Bohr angular frequency of the transition between these two levels ( h0 = Eb Ea ) and we take the energy reference in the middle of the two levels. 1.1 At t, the system is supposed to be described by the ket: | (t) = ca (t)|a + cb (t)|b , (1)

where ca (t) and cb (t) are complex. If the atom is isolated, what is the equation governing the state | (t) ? How do ca (t) and cb (t) evolve with t? 1.2 Rather than describing the system by = ca |a + cb |b , we can choose to describe it by the operator = | |. (a) Show that is a projection operator. (b) Express the elements of the matrix representing in the basis formed by the states |a and |b as a function of ca and cb . What is the trace of ? (c) Consider an observable A. Write the average value A of A in state | as a function of ca , cb and the matrix elements of A in the {|a , |b } basis. Generalise this expression to a system of arbitrary dimension and basis {|un }. Show that A = Tr{ A}.

Mixed Case

We now suppose that the system is in a statistical mixture of states | with respective (classical) probabilities p . The density operator is then dened by = p | |.

2.1 Is Hermitian? What is the trace of ? 2.2 What is the equation of evolution of ? 2.3 What is the average value A of A for the system described by ? 2.4 What is the physical meaning of the matrix elements ij of in the {|a , |b } basis? 2.5 Consider a system in which the atoms are in superpositions of levels |a and |b with equal weights but with a relative phase between the parts of the wave function lying in levels |a and |b uniformly distributed between 0 and 2 . What is the density matrix of the system? Is 2 equal to ?

Rabi Oscillations

We again suppose that the system can be described by a pure state like in eq. (??). We now suppose that the atom is excited by a monochromatic eld linearly polarised along Ox given by E = Ex ux = (Aeit + c.c.)ux , where ux is a unit vector along Ox. The action of the electromagnetic eld on the atoms is described by the electric dipole Hamiltonian HI = d E = dx Ex . In the {|a , |b } basis, the operator dx reads: 0 d dx = . d 0 We suppose that d is real. 3.1 We write the state of the atom as: | (t) = a (t)eit/2 |a + b (t)eit/2 |b . We note: 1 =
2

(2)

2dA , h = 0 , = |1 | + .
2 2

(3) (4) (5)

What are the equations of evolution of a and b ? 3.2 What are the fast and slow terms in these equations? The quasi-resonant approximation consists in keeping only the slowly varying terms. 3.3 Suppose the system is initially (at time t = 0) in state |a . What is the state | (t) ? What is the probability Pb (t) for the system to be in state |b at time t. Plot Pb (t) versus time. How does the amplitude of these Rabi oscillation evolve with ? 3.4 Why are such oscillations never observed when a classical light source shines an absorbing medium?

 % % % ! n( z ' , t ' ) = n 2 ( z ' , t ' ) n1 ( z ' , t ' ) # % & % ' % (! !

! "# $ % )

"

"

2
& %) &
1 I I + = NI c t ' z '

% &

% )

& ! " %

# $

&

&
z = z' t = t ' z ' / c

, ! % )

# ) ! " % & &

./ ! "

N 1 NI = t ' h
0 , , ) ) . ! " ) % %

N 2 NI = t ' h
. # 1# " +

"!

J ( z, t ) =

I ( z , t ) dt

I ( z, t ) = J ( z, t )

$ %)
N2!

JS =

h #$ 2

&

u ( z , t ) = J ( z , t ) / J S #, Nt #

!"

N 2 ! t

%
N 2 ( z, t ) =

.
Nt + K ( z ) exp( u ( z , t ) ) 2 u ( z , t = ) # 0

&

3 " %& 45" ! % 0 ) & ) %

N 2 ( z , t = ) ! N t !

& !
N 0 # ,

& )

+ u ( z , t = ) = 0

0 = N 0 #

62# $ +

u ( z , t ) = 0 [1 exp( u ( z , t ) )] z

. 72 8

! "

%
G 0 = exp( 0 L ) #

% %
exp(u 2 (t ) )

&

#$
exp(u1 (t ) )

% #$ 72# %

"

!
I 2 (t ) = I 1 (t )

+
G 0 exp(u1 (t ) ) 1 + G 0 [exp(u1 (t ) ) 1]

, , # , &
/

"

7
G 0 J 1 << J s #

$# " #

% " " # : 1/ % #; % %: # 1

J 1 >> J s #

& #, %

1/! ! 9.2!

1 ;" ! %

% & &

" %

" #

< ;

&

& %#

TD 3 : Laser oscillation build-up analysis

We will address the laser oscillation build-up mechanism: how does the photons number and the inversion population evolves when the laser is first turned on? First, we consider a He-Ne laser, following a sudden initial turn-on using an electro-optic modulator (for example) inside the cavity. We register the emitted optical signal as a function of the time parameter t. The resulting curve is represented by the figure 1. We note by I(t) and Isat respectively the optical intensity and the saturation intensity.

Figure 1: He-Ne Laser Intensity as a function of time t.

The photons number, F, and the inversion population number, !N, rate equations are reminded below:

dF F =" + # F !N dt $ cav

Eq: 1

d!N 1 = " (!N " !N 0 ) " # F !N dt $

Eq: 2

1. Derive the steady state solutions for I and !N as a function of Isat, !N0 and the inversion population number at threshold, !Ns. 2. We first study the non-saturated amplification regime. This one corresponds to the linear part of the experimental curve of figure 1. At t = 0 , the inversion population number is equal to !N 0 and I = I O << I sat .

2.1. Proof that the optical intensity I(t) builds up exponentially with time starting from an initial noise level I 0 . Express the build up time constant, !l, in terms of !cav !N 0 and the pumping rate r = . !N s 2.2. Assuming that !cav = 100 ns, measure the pumping rate, r, corresponding to the curve of figure 1. 2.3. The total intensity builds up to the steady state oscillation level Ist. Express then measure the build-up time Tb. 3. The laser gain will saturate immediately with increasing light intensity I (t ) ( see the asymptotic behaviour of the experimental curve when increasing t ). 3.1. Can we consider that !N (t ) follows instantaneously the variations of the optical intensity I (t ) when assuming that (r " 1) << the considered laser, when ! = 500 ns ?

! cav ? Is this assumption correct for !

3.2. Taking into account the previous assumptions, show that the inversion population !N (t ) is expressed by the following equation:

& I (t )# (N (t ) = (N 0 $1 ' ! % I sat "

3.3. Using the previous result, express

Eq: 3

dI (t ) as a power of I (t ) . dt

3.4. Using the identity given by the equation 5, show that I (t ) reads:

-t * I st . I 0 . exp+ +/ ( ( , l) I (t ) = & & - t * ## $ I st + I 0 . $exp+ +/ ( ( ' 1! ! l $ , ) "! % % "

Eq: 4

1 1 &1 1 # = $ + I (b ' I ) b % I b ' I ! "

Eq: 5

3.5. Does the equation 4 fits the experimental results of figure 1? 3.6. Express the time delay following gain turn-on needed to reach, half of the steady-state oscillation level. Compare it to the expression found in 2.3

TD 4 : Observation and calculation of laser frequency pulling Laser Physics, Class Exercise #4
We intend, in the following exercise, to study the exact oscillation frequency of a laser. Laser oscillation, normally, occurs in only a few preferred longitudinal and transverse modes of the laser cavity. The exact oscillation frequency of a laser will, however, be shifted by a small amount from the resonance frequency of the cold cavity (the cavity without the gain medium) because of some dispersion effects associated to the gain medium. In practice, its quite difficult to study the absolute frequency of a laser oscillator to very high precision. However, its much easier to measure relative laser frequencies by observing the difference or beat frequency between two different laser signals or two longitudinal modes. Let us consider a 30-cm long He-Ne laser sustaining the oscillation of multiple axial modes. We detect the laser beam with a high-speed photodiode and then look at the obtained electrical signal with a radio-frequency (RF) spectrum analyser. The experimental set-up is depicted in the following figure:

437.5 MHz
Multimode He-Ne Photodiode

1 MHz
RF spectrum analyser
Figure 1 : Measurement system for observing the multimode He-Ne RF spectrum

1- Give a simple explanation (with a scheme for example) for the existence of 2 peaks. Lets now look at how these dispersion effects can be calculated. First, lets consider a linear cavity of length Lcav/2. R1 and R2 are the mirrors reflectivities and ! holds for the other optical losses. The intracavity electromagnetic field, described by a plane wave, reads: E (z, t ) = !e"iwt + ikz + c.c. 2- Remind the equation equivalent to the fact that the field is equal to itself after propagation through one cavity round trip. 3- The wave number, k, is complex and reads: ' '' % " ! at n"( ! at # k= 0 & 1 + + i 2 # c0 & ' 2n0 $ c0 n0
' '' where ! at and ! at are, respectively, the real and imaginary part of the susceptibility ! at .

' Express the frequency, ", of the laser as a function of " at ! q , n0, c0 and "q the cold cavity resonance frequency of order q.

( )

4- Plot the order q as a function of the laser frequency "q ( in a cold cavity) and the laser frequency " (including a dispersive active medium). Show that the cavity resonance frequencies are pulled when including the optical gain medium.
' 2 5- We remind that ! at . Express the laser frequency " as: ! 2n0 " = " q + !" q

!" q is called the frequency shift amount.

On which experimental parameters depends the frequency shift amount? 6- Compare the dispersion effects of an amplifying atomic transition to those of an absorbing atomic transition. 7- Here, we consider optical frequencies where the dispersion region is approximately flat. 7.1. Show that the axial-mode spacing between successive modes remains nearly constant but is decreased by the mode-pulling effect when compared to the cold cavity case. 7.2. Does the experimental result of Figure 1 correspond to the flat region of the He-Ne atomic transition?

8- Lets consider a Lorentzian atomic transition at a central frequecy "0 and a FWHM "! . '' In this case, we remind that ! at exhibits a Lorentzian profile. The associated dispersion
' profile ! at reads:
' (! ) = $ " at

2 Express the frequency pulling amount !" q as a function of ! q , ! 0 , "! , c0, n0 and the optical
gain per unit length " ! q . 9- Show that the gain medium shifts the laser frequency towards the gain maximum (i.e the center line "0). This is the so-called frequency pulling effect. 10- In which case, the pulling effects become large? 11- As a numerical example, we consider a He-Ne Laser with 4% power gain per round trip and 3 axial modes within the atomic linewidth. Calculate the pulling amount of the first axial mode compared to the axial mode spacing "! q .

! $ ! 0 '' " at (! ) #!

( )

! "# %&' ( , *0 * ! ! !)* + . ! "* /## - * ! /## * *0 3'2#,4 3"2"*

1 .

2 '## - *

Input mirror M1 Pulsed pump

Output mirror M2 Output (800 nm)

I
Ti:sapphire L

'# - * * + ( 0 1 , ) 6'6*
Intensity Pump pulse (10 ns)

'* 5 * 1 ,. . ,

6"6)* 5

1 Laser pulse Build-Up time B 2

Time

7 !

, ! *0 8 *+ ( ')* . ! - ,
1/2

! * * 2 2 /## * * ! ! . . '0* *
'

0 0 9

$ - ,. ,

!. - *

3 + 2# 1 1 1 1 9 . < ; + 0 ? ' + ,. 1

! -

8 !

* : *

2 * ; () - - ! < =2&> 2"# , . * . ; ( )< . . *+ ,. . * ! < + -

; ()

; ()

? *@

! ! !

1 *

1 1 1

@ 0 +

0A ; . 0

( , -, *+ !

0BB0 ) *

I 0 n max 2 d / , .

C 1 ,

- ! ! . 1 1 1 1 H 1 1 1 + H + . < + =D * 9 . ; . !3'2#*4, 2'## G ! ( 0

*+ . ! ! 0 )* 0

! -

0#< + E2'/F"#1'# , ; , =D
'

. , 2;

A;

E, , 3", 3'< ! 2%, 0#2#*/+ * 1'* *

, 2"#

, 3"2",

! ! ? *+

! . ( I

. ')*

! ! ! !

- ! !

< < -

2/2

Excercise # 7 : Laser linewidth characterization in a delayed self-heterodyne interferometer

Intensity and phase dynamics of an optical source are key characteristics in determining the performance of optical systems. We will focus, in the following exercise, on the measurement of the laser linewidth (see the following figure). One dominant cause of spectral broadening in single-longitudinal-mdoe laser is phase noise. Random phase noise is created when spontaneous emission, originating in the laser cavity gain media, changes the phase of the laser frequency. On the contrary to intensity fluctuations, optical phase noise is not detectable with a simple optical power detection. When used in an interferometer, phase modulation (or noise) is converted into intensity modulation (or intensity noise). This feature will be deeply developed as a method to characterize laser linewidth. Measurement assumptions and technique: The laser is assumed to operate in single-longitudinal mode. The laser lineshape typically has a Lorentzian shape. The measurement technique is shown in Figure.

Incident light is split into two paths by a fibred interferometer. An acousto-optic frequency shifter is inserted in the first arm of the interferometer. It provides a 80 MHz frequency shift of the light in arm 1 with respect to the other arm. The arm 2 of the interferometer contains a temporal delay ! d . The optical signal is then sent to a photodiode. The resulting electrical signal is analysed with an electrical spectrum analyser. At the laser output, the electrical field can be expressed as:

E (t ) = E0e( j 2"# 0 t + j! (t )) + cc,

equation 1

Where E0 is the light field amplitude which we suppose constant. ! 0 is the average optical frequency. ! (t ) is the optical phase noise. We wuppose that the laser phase noise is the white quantum noise. In this case, the auto-correlation of the frequency error is given by:

# (t + ! )# (t )

!c

" (! )

equation 2

! c is the laser coherence time. The coherence time of a laser is a measure of the
spectral purity of the laser frequency over time. Coherence time is reduced by random events, such as spontaneous emission in the laser cavity. Coherence time varies inversely with laser linewidth, "! . Its defined for spectra with Lorentzian lienshapes as:

#c =

1 2"$!

equation 3

1-Express the electrical field at the output of the interferometer. The frequency shift of the acousto-optic is noted by the frequency f AC . 2-Give the expression of the optical intensity incident on the photodetector, I(t). In the following part, we intend to obtain the spectrum of this signal. We need to calculate the auto-correlation function of I(t) and apply the Wiener-Khintchine theorem. The auto-correlation function is given by:

RI (! ) = I (t )I (t + ! )
Where the brackets denote ensemble averaging. 3- Demonstrate that the auto-correlation function R I reads :
4 4 RI (! ) = 4E0 + 2E0 cos H (t ,! ) cos(# 2"f AC! )

equation 4

equation 5 equation 6

where

H (t ,! ) = " (t + ! ) # " (t ) # " (t # ! d + ! ) + " (t # ! d )

We suppose that the phase jitter defined by: $! (t ," ) = ! (t + " ) # ! (t ) is a zero-mean stationary Gaussian process. Then we have the following relations:

cos('* (t ,) ))

& 1 # ( '* 2 (t ,) ) ! $ " = e% 2

cos(H (t ,( )) 4-Using the equation 2, demonstrate that:

& 1 # ' H 2 (t ,( ) ! $ " = e% 2

$ % if % < % d !2 ! %c 2 equation 7 H (t ,% ) = # % !2 d if % > % . d ! " %c 5- Finally, by taking the Fourier transform of equation 5 ( and using equation 7), the power spectral density of the detected signal reads as:
4 4 S I (w) = 4 E0 * (w) + E0 e '

(d (c

(* (w ' 2)f AC ) + * (w + 2)f AC )) +

equation8

2 (w ' 2)f AC )2 1+(c

(c e )

'

(d (c

& (d # ( ( ) ) sin w ' 2 ) f ( ( AC d $e c ' cos((w ' 2)f )( ) ' ! AC d $ (w ' 2)f AC )( c ! % "

Comment the different contributions to the previous expression. 6- The following spectra are measured using the experimental set up depicted by A:TRACE219.CSV (E4402B) Figure 1.
-80 -85 -90 -95 -100 -105 -110 40 60 80 100 120
-30

Signal detecte (dBm)

-40

Signal (dBm)

-50

-60

-70

-80 77 78 79 80 81 82 83

Frquence (MHz)

Frequence (MHz)

Compare the two spectra and indicate which one is obtained for a delay time much smaller than the laser coherence time. Deduce the experimental conditions needed for a direct measurement of the laser linewidth. 7-Simplify the equation 8 for the obtained experimental conditions.

" "

! !# &

"

$%

A1 (t ) = e 1t

g l Dg

A2 (t ) = e 2t

e jM cos m t
* +

A3 (t ) = e 3t

'

( !

"

" )

"

, -.! 1 !# # " 2 " !3 ) " "

$( //% 0 2

" " "

" " 2

+ jM cosmt

jM sin mt

!4

3 ! "

"

"

"

" ) !5 " 4 2' -.1!

1/2

" "

! "

"

a(t ) = e (t )

e M (1cos mt ) ! #

"

" " -.2 0 ' &

" " 5 !, 9 2 $ " "

"" !7

! )

"

" 8 " 8 % + -:;<=. *' > + -:;<:.. ' > *-:?@

= M 4 2 g 1
1/ 2

" ) -?@ ! A

m g !

" "" "

"

%. -.2 #

! 2

2/2

Dispersion and pulse characterization

Laser physics Exercise class #10

1 Influence of dispersion on a chirped Gaussian pulse We consider a chirped Gaussian pulse propagating in a dispersive medium with group-velocity dispersion coefficient 2. The pulse entering the medium is written as:
E (t , z 0) exp 1 iC t 2 2 2 t0

Where C is the chirp parameter and t0 is the half-width at 1/e. 1 What is the spectrum of this pulse? How does the spectral width compare to the unchirped Gaussian pulse case? 2 The pulse is launched in a dispersive medium. What is the temporal width of the pulse after propagating through a length z? 3 Study the evolution of the pulse width as a function of z when C and 2 have the same or opposite signs. 4 If C2<0, what is the distance that minimizes the pulse width? What is the corresponding pulsewidth. Does the result make sense with that of question 1? 5 Consider a Gaussian pulse with t0=100 fs, C=3 at 1550 nm. What length of optical fiber with 2= -0.02 ps2.m-1 should be used to compress this pulse, and what is the minimum pulsewidth?

2 Characterization of ultrashort pulses Consider the unbalanced Michelson interferometer shown below, with an adjustable delay between both replicas of the pulses . We want to use it to characterize ultrashort pulses.

1 Show that the physically interesting part in the detected signal can be written as:
( ) E (t ) E (t ) dt ,

Where E(t) is the incident electric field. We can measure this quantity as a function of to characterize the pulse.

1/2

2 Calculate the Fourier transform of (). What kind of information do we get from this measurement? 3 As an example, consider the case of exercise 1. What would be the result of our measurement on the pulse at the input of the dispersive medium, and at the output of the medium? Conclude.

Suppose now that we want to measure a Gaussian pulse of width T with an intensity autocorrelator, shown below

In this case, the complex envelope of the second harmonic generation field (represented by the green beam) is proportional to E(t ) E(t ) . 4 Show that the measured signal is now given by
A( ) I (t ) I (t ) dt ,

Where I is the intensity of the incident pulse. 5 We want to measure a Gaussian pulse of width T with an intensity autocorrelator. This pulse intensity can be written as I (t ) exp
t2 . What is the autocorrelation signal as a function of the delay introduced T2

between the arms of the autocorrelator ? 6 How can we deduce the pulsewidth from this measurement? If we measure an autocorrelation width of 140 fs, what is the corresponding pulsewidth ? Can this technique allow us to distinguish a pulse from the same pulse broadened in a dispersive medium? 7 What is the influence of pulse shape on the pulsewidth measurement (square, sech, lorentzian, etc). 8 The experimental setup uses noncollinear SHG. What would happen in the case of a collinear SHG setup ?

2/2

Stability of a cavity containing a gain medium with an index gradient

Laser Physics Class Exercise #12 9/12/09

We consider an optical cavity between two concave mirrors, with radius of curvature R1 and R2, separated by a distance L (Ri > 0). 1-Show that this cavity is equivalent to the one represented below and composed of plane mirrors M1, M2 and lenses of focals fi= Ri. Express the matrix T for propagation from M1 to M2 as a function of gi=1-L/Ri and L. We remind that the cavity is stable when 0<g1g2<1. L

z M1 f1 f2 M2

2-We insert inside the cavity an amplifying bar which is uniformly pumped but only cooled in the suburb, leading to an inhomogeneous repartition of temperature and then the optical index. As a first approximation, we consider that the index variation makes the gain medium equivalent to a convergent lens of focal length f. We intend to study the stability of the cavity represented below, considering the propagation matrix from M1 to M2. L1 L2

z M1 f1 f f2 M2

-Calculate as a function of R1, R2, L1, L2 and f the three elements of the new matrix T defining the new parameters g1, g2 and L. Verify the expected expressions of g1, g2 and L when considering the simple cavity (without the gain medium). 3-We intend now to determine more precisely the effects of the index variations. The index depends on r, the distance of the optical ray to the propagation axis z (in a plane perpendicular to z). Demonstrate that the quantity n( r ) cos( ) is conserved during the propagation of an optical ray in the gain medium. Deduce, in the paraxial approximation and when supposing that n(r ) is given by:

that the function r ( z ) verifies the following equation:

1 1 n(r ) = n0 n2 r 2 and n2 r 2 pp n0 2 2

d 2r dz 2
cos Tb = n0 n2 sin

n2 r (z ) n0

4-Deduce that for a gain medium of length l ; the propagation matrix between the planes P and P writes as:

1 sin n2 n0 n 2 with = l n0 cos

P is the plane of entrance of the optical ray in the gain medium. Pcorresponds to the plane at the output of the gain medium. 5- In the following questions, we suppose that pp 1 . Demonstrate that the propagation matrix of the gain medium is equivalent to the one corresponding to the following system: h h

z P f P

with a focal length f and a distance h to be determined. 6- What are the values of L1, L2 and f that should be used with the results of Q-2 in order to obtain the condition of stability of the cavity represented below? d1 d d2

z M1 f1 f2 M2

1
An illustration of spatio-temporal couplings of ultrashort laser beams

The ultrafast lighthouse eect

CONTEXT

The process of high-harmonic generation, through the non-linear interaction of an intense laser eld with a medium, generally leads, in the time domain, to the generation of a train of attosecond pulses. An attosecond pulse is emitted every half optical period or every optical period (typically 2-3 fs), depending on the symmetry of the medium. Isolated attosecond pulses are much more convenient for applications, for instance for pump-probe experiments. In the past 10 years, many eorts are therefore been devoted to the development of interaction schemes that lead to the generation of isolated attosecond pulses, rather than such trains. This has in particular lead to the development of laser sources delivering intense pulses which only consist of one to two optical periods.

THE ULTRAFAST LIGHTHOUSE EFFECT

In this exercice class, we will consider a recently developed approach to this problem, which can be used even with many optical-period long laser pulses. The basic idea is to generate the train of attosecond pulses in such a way that all the pulses of this train propagate in slightly dierent directions, so that they produce angularly-separated beams as they diverge away from the generation medium. This is called an attosecond lighthouse. Each attosecond pulse is generated in the direction of propagation of the laser eld at the time of generation, i.e. along the normal to the instantaneous laser wavefront. To produce an attosecond lighthouse, one thus needs a light eld at the point of interaction with the medium, such that it wavefronts rotate in time. The rotation has to occur on the femtosecond time scale. In this class, we will see how to induce such a wavefront rotation, called the ultrafast lighthouse eect.

DEFINITIONS
dimension for simplicity). We assume that this eld can initially be decomposed as :
Exi t = f (t)g (xi )eiL t

We consider a femtosecond laser beam, of central wavelength L . The laser eld is written Exi t = h(xi , t), where t is time and xi is a spatial coordinate transverse to the propagation direction z (we neglect the second transverse (1)

where L is the central angular frequency of the laser pulse. The eld is then said to exhibit no spatio-temporal coupling : the temporal properties of the light pulse are independent of space.

PULSE FRONT TILT

This laser beam is propagated across a prism, made of BK7 glass.
Question 1

Considering that the group velocity vg and phase velocity vp of light are dierent in glass, and that dierent parts of the beam cross dierent thicknesses of glass, describe the relative position of the laser pulse wavefronts and envelop after this prism (assuming a thin prism).
We now decompose the laser eld as :
Exi t = f (t xi )g (xi )eiL t

(2)

where is called the pulse front tilt parameter, which has the dimension of a time divided by a length (typically expressed in f s/mm). This eld exhibit a spatio-temporal coupling : in this case, the arrival time of the pulse varies with the transverse spatial coordinate xi .

2
Question 2

Calculate for a laser pulse centered at 800 nm, that has traveled through a thin BK7 prism with a 1o apex angle (at 800 nm, vg = 1.964 108 m/s and vp = 1.988 108 m/s in BK7).

ANGULAR DISPERSION
Glass is a dispersive medium (leading to vg = vp ). As a result, the dierent frequency components of the ultrashort laser pulse are deected in dierent directions (), as the laser beam crosses the prism. The quantity = d/d is called the angular dispersion. We now calculate the relationship between pulse front tilt and angular dispersion.

Calculate the spatially-resolved complex frequency spectrum Exi (function of xi and ) of Exi t . Analyze the result physically.
Question 4

Question 3

Calculate the relationship between and .

SPATIAL CHIRP AT FOCUS

In order to generate harmonics and attosecond pulses, very high laser intensities are required, and to this end the laser beam needs to be focused on the target. To this end, a perfect focusing optics of focal length f is used. Assuming that the laser beam has a nite pulse front tilt before focusing, we now determine the laser eld conguration at focus. The transverse spatial coordinate at focus is called xf .
Question 5

Considering the relationship between angular dispersion and pulse front tilt, discuss qualitatively what happens at the focus.
The eld Exf (function of xf and ) at focus can be deduced from the 2D Fourier-transform Ek (function of k and ) of Exi t , using k = kL xf /f , with kL = 2/L the central wave vector of the laser beam.
Question 6

Starting from Exi (calculated in question 3), calculate Ek . Deduce the eld at focus Exf , and analyze the result physically.

WAVEFRONT ROTATION
proles before focusing, i.e. We now wish to calculate the eld as a function of time t and position xf across the focus, i.e. Exf t (function of xf and t) instead of Exf . To simplify this calculation, we assume that the beams has Gaussian spatial and temporal

Exi t = E0 exp 2
Question 7

t xi i

x2 i 2 wi

exp (iL t)

(3)

Assuming the above expression for Exi t , calcule Ekt , and deduce Exf t (in this calculation, omit all terms that are independent of time or wavevector k, and concentrate only on the exponential that depend on these two variables). Discuss physically the obtained duration and focal spot size. Comment on the phase terms.

3
Question 8

Use the expresion obtained for Exf t , calculate the instantaneous wavefront direction at focus, using k /kL , with k = /xf , where is the phase of Exf t . Calculate the wavefront rotation velocity vr = d/dt. Show that it has a maximum as a function of , and determine the optimal value of . Explain this maximum physically.
Question 9
max From the expression obtained for vr , derive the condition that needs to be fullled on the ratio of the harmonic and laser divergence, n /L , for given laser beam parameters, to produce an attosecond lighthouse. Comment physically.