Lecture 5: Diffraction Grating

1 Diffraction Grating

Diffraction grating is basically a collection of equi-distant slits with opaque regions in between.
One can achieve that same by considering an optically thin plane glass plate on which a number
of parallel and equidistant lines are drawn. The transparent region of the glass plate act as
slit while the ruled lines separate these slits, which behave as opaque regions. Typically, there
are about hundred slits in each millimeter length of a diffraction grating. For generality, let us
assume that there are total N number of slits in a diffraction grating, such that width of each
slit if given by a and the distance between two consecutive slits is b (see Figure 1).

Figure 1: The setup of diffraction grating.

Let a parallel beam of light ray of wavelength λ falls on such a diffraction grating. At each
point of the slit the light source act as secondary wavelets and thus will interfere as they reach
the screen. Consider a particular slit denoted as P Q. Following the discussion of single slit
diffraction pattern, here also the phase difference between two light rays coming from two
different points of the slit separated by a distance x is given by,
2π
× Path difference = `x , (1)
λ

1
Optics Indian Association for the Cultivation of Science

where ` = (2π/λ) sin θ. It is advantageous to define the following quantity, d = a + b, in terms

of which the total amplitude at the point O on the screen is given by
Z a/2 Z d+a/2 Z (N −1)d+a/2
i(ωt−`x) i(ωt−`x)
Atotal = dx Ae + dx Ae + ··· + dx Aei(ωt−`x)
−a/2 d−a/2 (N −1)d−a/2
−i`a/2 −i`(d+a/2) −i`(d−a/2) −i`{(N −1)d+a/2}
i`a/2
− e−i`{(N −1)d−a/2}
 
iωt e −e e −e e
= Ae + + ··· +
−i` −i` −il
      
2 `a 2 `a 2 `a
= Aeiωt sin + e−i`d × sin + · · · + e−i`(N −1)d × sin
` 2 ` 2 ` 2
 
2 `a
= Aeiωt 1 + e−i`d + ....... + e−i`(N −1)d × sin
 
` 2
 −i`(N −1)d 
sin α iωt e −1
= Aa e −i`d
, (2)
α e −1
where, α ≡ (`a/2). Thus the total intensity at point O is given by
2
 
2 2 2 sin α 2 − 2 cos(N `d)
I ≡ |Atotal | = A a
α2 2 − 2 cos(`d)
2 2 2
 
2 2 sin α 1 − cos(N `d) 2 2 sin α sin (N `d/2)
=A a = A a
α2 1 − cos(`d) α2 sin2 (`d/2)
sin2 α sin2 (N β)
= A2 a2 2 , (3)
α sin2 β
where we have defined, β ≡ (`d/2). For N = 2, i.e., diffraction from double slit, we have
I = A2 a2 (sin2 α/α2 ) cos2 β, which coincides with the expression of total intensity derived in
the previous lecture. Hence we can conclude that the term (sin2 α/α2 ) yields the diffraction
pattern from a single slit, while (sin2 N β/ sin2 β) yield the interference pattern from N slits. We
will now characterize the location of maxima and minima arising out of the above diffraction
pattern from N slits.

• Principal maxima: If the slit width a is very small compared to the separation d
between successive slits, then α is very small and hence the variation of the factor
(sin2 α/α2 ) is small. As a result, the maxima is determined by the interference term
(sin2 N β/ sin2 β) alone. This factor will be maximum when β = mπ, where m =
0, ±1, ±2, . . .. Then the maxima will be proportional to the number of slits as in the
limit β → mπ, (sin N β/ sin β) = N . So the condition of maxima in this case will be

(a + b) sin θ = mλ . (4)

This is known as condition for principal maxima. At the location of principal maxima,
the intensity becomes I = N 2 (sin2 α/α2 ). Thus for large number of slits in the grating
and small α the intensity significantly increases, while if we deviate from small α, the
intensity decreases. Note that α = 0 is the central maxima of the diffraction pattern.

• Secondary maxima and minima: To find out the secondary maxima and minima, we
need to consider the interference term (sin2 N β/ sin2 β) more closely. The condition for

2
Optics Indian Association for the Cultivation of Science

extrema will be determined by taking derivative with respect to β, which yields,

dI
=0
dβ
2 sin(N β) cos(N β)N 2 sin2 (N β) cos β
=⇒ − =0
sin2 β sin3 β
2 sin2 N β
=⇒ [N cot N β − cot β] = 0 . (5)
sin2 β

Hence, for the condition of extrema, we have either (sin2 N β/ sin2 β) = 0 or, N cot N β =
cot β. To see which corresponds to maxima and which are the minima, we need to compute
the second derivatives of the intensity, yielding,
2
d2 I 2 2 sin N β 2 sin2 N β
−N 2 csc2 N β + csc2 β


2
= 2 (N cot N β − cot β) 2 + 2
dβ sin β sin β
2
2 sin N β 
2 (N cot N β − cot β)2 − N 2 (1 + cot2 N β) + 1 + cot2 β

= 2
sin β
2 sin2 N β  2 2 2 2 2

= 2 (N cot N β − cot β) − N cot N β + cot β − N + 1
sin2 β
d2 I  2 sin2 N β

=⇒ = (−N 2 + 1) < 0 . (6)
dβ 2 N cot β=cot β sin2 β

Clearly the condition N cot N β − cot β = 0 is a condition for maximum intensity. These
are known as “secondary maxima”. On the other hand, while sin N β = 0 (assuming
sin β 6= 0) i.e., N β = mπ, where m = ±1, ±2, . . ., we have zero intensity i.e., minimum
intensity. Hence, the minimum condition can be given in terms of wavelength as

N (a + b) sin θ = mλ . (7)

The ratio of secondary maxima and primary maxima is an important quantity, which can
be determined as follows (with I0 = A2 a2 )

sin2 N β sin2 α 
 
Ism = I0
sin2 β α2 N cot N β=cot β
sin2 α
 
1 
= I0 2 2 2

α sin β (1 + cot N β)  N cot N β=cot β
2 2
sin α N
= I0 2
α 2 sin β(N 2 + cot2 β)
1 1
= Ipm 2 2 = Ipm
N sin β + cos β2 1 + (N − 1) sin2 β
2

Ism 1
=⇒ = , (8)
Ipm 1 + (N − 1) sin2 β
2

where the intensity at principal maxima is defined as, Ipm = N I0 (sin2 α/α2 ). These points
have been depicted in Figure 2.

3
Optics Indian Association for the Cultivation of Science

Figure 2: Primary maxima, secondary maxima and secondary minima arising out of diffraction
grating have been depicted.

• Absent spectra: For mth order principal maxima we have the condition

(a + b) sin θ = mλ , (9)

where, m = 0, ±1, ±2, . . .. On the other hand, for the pth order diffraction minima, we
know

a sin θ = pλ . (10)

If these two conditions are satisfied simultaneously then mth order principal maxima of
interference will be absent from the fringe patten. From combining the above two criterion
we have the following condition for missing fringe

a+b m
= . (11)
a p

Thus if the ratio (a + b/a) is rational, then there will be missing principal maxima in the
diffraction pattern of grating.

• Ghosts: In a normal grating the rulings are generally not equally spaced. As a result
the grating gives a continuous background illumination, since the intensity never exactly
goes to zero. Another very common error in a diffraction grating is that the error in
equally spaced slits can be periodic. Due to this error, additional false lines appears in
the grating as maxima. These false lines, arising out of erroneous preparation of grating
are known as ghosts.

So far we have discussed various maxima and minima associated with the diffraction pattern.
We will now discuss various observables associated with the diffraction grating.

• Dispersive power: Dispersive power gives the information about how good the grating
is in separating away closely spaced wavelengths. It is defined as the rate of change of

4
Optics Indian Association for the Cultivation of Science

the angle of diffraction with the change in wavelength i.e., (dθ/dλ). If θ is the angle of
diffraction for rays of wavelength λ forming the mth order principal maxima, then we have

(a + b) sin θ = mλ
dθ m
=⇒ = . (12)
dλ (a + b) cos θ

Thus larger the value of m and smaller the value of (a + b), larger will be (dθ/dλ). Also
larger θ means smaller cos θ and hence larger (dθ/dλ). This means the further we are
from the central location, better is the dispersive power. Also gratings which are more
finer (i.e., have smaller (a + b) values) have higher dispersive power. In addition note
that for small θ, we have cos θ ≈ 1. Then we have (dθ/dλ) ≈ constant. Hence dθ ∝ dλ.
So the angular separation is same as the wavelength separation. These are called normal
spectrum.

• Resolving Power and Rayleigh criterion of resolution: Two equally bright spots
can be resolved by an optical system if the angular separation between them is such that
the central maxima due to one coincides with first minima of the other. This is known as
the “Rayleigh Criterion” for resolution. In this case, there will be a dip halfway between
them. In other words, the angular separation between the two principal maxima is equal
to half the angular width of either principal maxima.

• Resolving power of a grating: The resolving power measures ability to distinguish two
closely spaced spectral lines with wavelengths λ and λ + ∆λ. Then the angular separation
between the principal maxima formed by these two wavelengths of order m is given by,
m∆λ
∆θ = . (13)
(a + b) cos θ

If for wavelength λ, the principal maxima appears at, angle θ, such that N (a + b) sin θ =
N mλ. Then the next secondary minima for the same wavelength will appear at an angle
θ + δθ, such that, N (a + b) sin(θ + δθ) = N mλ + λ. This is because, for the next order
secondary minima, ∆(N β) = π, yielding, ∆(N d sin θ) = λ. Thus we obtain,

N (a + b) sin(θ + δθ) = N mλ + λ
sin(θ + δθ) N mλ + λ 1
=⇒ = =1+
sin θ N mλ Nm
sin θ + δθ cos θ 1
=⇒ =1+
sin θ Nm
1
=⇒ δθ = . (14)
N m cot θ
The two wavelengths will be said to be resolved by the grating, if the principal maxima
of one wavelength coincides with the secondary minima of the other. Thus we want the
secondary minima of wavelength λ to appear at θ + ∆θ, such that it will coincide with
the mth order principal maxima of λ + ∆λ. This requires, ∆θ = δθ and hence we obtain,
m∆λ 1
=
(a + b) cos θ N m cot θ

5
Optics Indian Association for the Cultivation of Science

1 m∆λ ∆λ
=⇒ = =
Nm (a + b) sin θ λ
λ
=⇒ = Nm . (15)
∆λ
Hence, the resolving power, defined as (λ/∆λ) increases as the total number of slits in
the diffraction grating N increases. The above expression can also be written down as,

λ N (a + b) sin θ W sin θ
= = (16)
∆λ λ λ
where, W = N (a + b) is the total width of the diffraction grating. Thus for a fixed W and
θ, λ/∆λ is fixed, so there is no improvement in the resolving power. On the other hand,
the dispersive power is proportional to 1/(a + b), meaning smaller (a + b) for a fixed N
indicates larger dispersive power but smaller resolving power.