Approximation Methods

For Solving the Schrodinger equation of multielectron systems

 Hamiltonian For Multielectron Systems (in a.u)
For Two electron and one nucleus (H-1 , He etc.) 1 r12
2
1 1 1 𝑧 𝑧 1
𝐻=− ∇2𝑎 − ∇12 − ∇22 − − + r1
2𝑚𝑎 2 2 𝑟1 𝑟2 𝑟12 r2
Assuming the nucleus is fixed we have
1 2 1 2 𝑧 𝑧 1 1
Ignoring this led to the simplest
𝐻 = − ∇1 − ∇2 − − + =𝐻 1 +𝐻 2 + Independent electron approximation
2 2 𝑟1 𝑟2 𝑟12 𝑟12
𝐻 =𝐻 1 +𝐻 2 (38% error)
For n electron and one nuclei Molecules
𝑛 𝑛 𝑛−1 𝑛
1 1 𝑧 1 For Two electron and two nuclei (in a.u)
𝐻=− ∇2𝑎 − ෍ ∇2𝑖 − ෍ + ෍ ෍
2𝑚𝑎 2 𝑟𝑗 𝑟𝑖𝑗 1 1 1 1 𝑧𝑎 𝑧𝑏 𝑧𝑎 𝑧𝑏 1
𝑖=1 𝑗=1 𝑖=1 𝑗=𝑖+1
𝐻= ∇2𝑎 ∇2𝑏 − ∇12 − ∇22 − − − − +
2𝑚𝑎 2𝑚𝑏 2 2 𝑟1 𝑟1 𝑟2 𝑟2 𝑟12
Many Body Problem : 𝑛
1
෍ For n electrons and k nuclei (in a.u)
Even three body problems are 𝑟𝑖𝑗
𝑗≠𝑖 𝑘 𝑛 𝑘 𝑛 𝑛−1 𝑛
Analytically very difficult to solve. 1 1 𝑧𝑗 1
𝐻 = −෍ ∇2𝑖 − ෍ ∇2𝑖 − ෍ ෍ +෍ ෍
2𝑚𝑖 2 𝑟𝑖𝑗 𝑟𝑖𝑗
𝑖=1 𝑖=1 𝑗=1 𝑖=1 𝑖=1 𝑗=𝑖+1
Electron-Electron interactions grow more rapidly.
 Methods of Approximation
Variational Approach
Start with the Time independent Schrodinger equation 𝐻ψ = 𝐸ψ
Left multiply with ψ* and integrating over its domain we have
‫ ׬‬ψ∗ 𝐻 ψ 𝑑𝞽
𝐸= (Expectation Value Postulate)
‫ ׬‬ψ∗ ψ 𝑑𝞽
𝐺𝑢𝑒𝑠𝑠 𝑎 𝑡𝑟𝑖𝑎𝑙 𝑤𝑎𝑣𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝞥 𝑜𝑏𝑒𝑦𝑖𝑛𝑔 𝑡ℎ𝑒 𝑏𝑜𝑢𝑛𝑑𝑎𝑟𝑦 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛𝑠 Our treatment science will be adequate,
‫𝞽𝑑 𝞥𝐻 ∗𝞥 ׬‬ if it achieves the amount of precision
𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒 𝑖𝑡𝑠 𝑒𝑥𝑝𝑒𝑐𝑎𝑡𝑖𝑜𝑛 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑒𝑛𝑒𝑟𝑔𝑦 𝐸 = which belongs to its subject matter.
‫𝞽𝑑 𝞥 ∗𝞥 ׬‬
𝐼𝑛𝑡𝑟𝑜𝑑𝑢𝑐𝑒 𝑣𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟𝑠 𝑎𝑛𝑑 𝑜𝑝𝑡𝑖𝑚𝑖𝑧𝑒.
𝐶𝑜𝑚𝑝𝑢𝑡𝑒𝑑 𝐸𝑛𝑒𝑟𝑔𝑦 𝑤𝑖𝑡ℎ 𝞥 𝑖𝑠 𝑎𝑙𝑤𝑎𝑦𝑠 ℎ𝑖𝑔ℎ𝑒𝑟 𝑡ℎ𝑎𝑛 𝑡ℎ𝑒 𝑡𝑟𝑢𝑒 𝑒𝑛𝑒𝑟𝑔𝑦.
Better to use a crude approximation and
Perturbational Approach know the truth, plus or minus Δx percent,
than demand an exact solution and know
Identify a solved system which is close to the problem system
nothing at all.
𝐻 = 𝐻0 + λ𝐻′
𝐻0 The Hamiltonian of a system whose solutions are known. An easily understood, workable falsehood
𝐻′ The perturbation term, need to be very small for efficiency. is more useful than a complex
incomprehensible truth.
Λ is the handle to switch the perturbation smoothly from off to on.
E𝑛𝑒𝑟𝑔𝑖𝑒𝑠 𝑎𝑛𝑑 𝑊𝑎𝑣𝑒𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛𝑠 𝑜𝑓 𝐻 𝑎𝑟𝑒 𝑒𝑥𝑝𝑎𝑛𝑑𝑒𝑑 𝑖𝑛 𝑇𝑎𝑦𝑙𝑜𝑟 𝑠𝑒𝑟𝑖𝑒𝑠
Improve solutions with the increasing order of perturbational correction.
 Mathematical Preliminaries
Hilbert Space
Hilbert space H - generalization of Euclidean space to n (finite or infinite) dimensions.
Generally complex vector space but with a positive definite inner product (dot
product) that allows the measurement of angles and lengths
Hilbert spaces form a complete set, with addition, Ψ1 , ψ2 ∊ 𝐻, Ψ1 +ψ2 ∊ 𝐻
The Vectors can also be multiplied by complex numbers: 𝑎 ∊ ℂ, Ψ ∊ 𝐻, 𝑡ℎ𝑒𝑛 𝑎Ψ ∊ 𝐻

Inner Product
The inner product of two wave functions: (𝞥, ψ) = ‫ 𝑟 ∗𝞥 ׬‬ψ 𝑟 𝑑 3 𝑟
Linear in its second argument 𝞥, C1 ψ1 + C2 ψ2 = C1 𝞥, ψ1 + C2 𝞥, ψ2
and anti-linear in the first C1 𝞥1 + C2 𝞥2 , ψ = 𝐶1∗ 𝞥1 , ψ + 𝐶2∗ 𝞥2 , ψ
It is also positive definite ψ2 = ψ, ψ ≥ 0 𝑢𝑛𝑙𝑒𝑠𝑠 ψ 𝑟 = 0
Also (𝞥, ψ) = (ψ, 𝞥)∗ and (𝞥, ψ)(ψ, 𝞥) ≤ (ψ, ψ)(𝞥, 𝞥)

Dirac Bra-Ket Notation Hilbert Von Neumann Dirac

The wave function ψ (some state in the quantum system) of the complex vector space is denoted as |ψ > (Ket)
A linear functional that maps each of this vectors into a complex plane (convectors) is denoted as < ψ| (Bra)
< 𝞥| ψ > = න 𝞥∗ 𝑟 ψ 𝑟 𝑑 3 𝑟 < 𝞥| 𝐴 |ψ > = න 𝞥∗ 𝑟 𝐴 ψ 𝑟 𝑑 3 𝑟
 Variational Method : The Recipe
Step 1: Study the system and its boundary
conditions Linear
2
Step 2: Guess a Trial Wave function 𝞥. Introduce Variable parameters 𝐶1 𝑥 𝐿 − 𝑥 + 𝐶2 𝑥 𝐿 − 𝑥
Step 3: Calculate the Expectation value of Non-Linear
𝛼
Energy ℰ associated with 𝞥 Optimize the parameters 𝑥 𝐿−𝑥
for minimizing energy
𝞥𝐻𝞥 δ𝐸 More the parameters,
ℰ= Evaluate Ci by setting =0 more the flexibility
𝞥𝞥 δ𝐶𝑖
more complex the problem
Step 4: The computed Energy is the upper more accuracy?
bound for the ground state Energy 𝐸𝑔
Example : Particle in a 1-D box
Boundary conditions : The wavefunctions should vanish at both sides of the wall (x=0 and x =L)
𝞥 = 𝑥 𝐿 − 𝑥 𝑓𝑜𝑟 0 < 𝑥 < 𝐿
Expectation value of <H> for the trial wave function
ℎ 2 d2
𝑥 𝐿 − 𝑥 − 8π2𝑚 2 𝑥 𝐿 − 𝑥 ℎ2 𝐿 2 d2
d𝑥 − 8π2𝑚 ‫׬‬0 𝐿𝑥 − 𝑥 2 (𝐿𝑥 − 𝑥 2 ) 5ℎ2 ℎ2
ℰ= = d𝑥 = 2 2 = 0.1266515
𝑥 𝐿−𝑥 𝑥 𝐿−𝑥 𝐿 2 2 4π 𝑚𝐿 𝑚𝐿2
‫׬‬0 𝐿𝑥 − 𝑥
The accidental similarity of the sin (x/l) with 𝑥 𝐿 − 𝑥 resulted in a very good agreement, but it is mere luck.
 Variational Approximation : The Theorem and its Proof
For a time independent H with the lowest Eigen value (ground state energy) E1, any well-behaved
function 𝞥 in the same domain (of the coordinates of the system’s particles) that satisfies the
boundary conditions will have its expectation value of energy ℰ ≥ Eg
𝞥𝐻𝞥
ℰ= ≥ 𝐸𝑔
𝞥𝞥 𝑛
Proof: 𝞥 can be expanded in terms of ψ𝑖 , the complete, 𝐶𝑖ψ𝑖 𝐻 𝐶𝑗ψ𝑗 = 𝐶𝑖∗ 𝐶𝑗 ψ𝑖 𝐻 ψ𝑗
𝞥 = ෍ 𝐶 𝑖 ψ𝑖 = 𝐶𝑖∗ 𝐶𝑗Ej ψ𝑖 ψ𝑗 ൿ
orthonormal set of stationary-state eigenfunctions of Hn,
𝑖=1
Energy expectation value associated with 𝞥 is 𝐶𝑖ψ𝑖 𝐶𝑗ψ𝑗 ൿ = 𝐶𝑖∗ 𝐶𝑗 ψ𝑖 ψ𝑗 ൿ
σ𝑛𝑖=1 𝐶𝑖ψ𝑖 𝐻 σ𝑛𝑖=1 𝐶𝑖ψ𝑖 𝐶1ψ1 + 𝐶2ψ2 + ⋯ . +𝐶𝑛ψ𝑛 𝐻 𝐶1ψ1 + 𝐶2ψ2 + ⋯ . +𝐶𝑛ψ𝑛 ψ𝑖 ψ𝑗 ൿ = δ𝑖𝑗
ℰ= =
σ𝑛𝑖=1 𝐶𝑖ψ𝑖 σ𝑛𝑖=1 𝐶𝑖ψ𝑖 ൿ σ𝑛𝑖=1 𝐶𝑖ψ𝑖 σ𝑛𝑖=1 𝐶𝑖ψ𝑖 ൿ
𝐶1ψ1 𝐻 𝐶1ψ1 + 𝐶1ψ1 𝐻 𝐶2ψ2 + ⋯ + 𝐶2ψ2 𝐻 𝐶1ψ1 + … 𝐶𝑛ψ𝑛 𝐻 𝐶𝑛ψ𝑛
=
𝐶1ψ1 𝐶1ψ1 ۧ + 𝐶1ψ1 𝐶2ψ2 ۧ + … + 𝐶2ψ2 𝐶1ψ1 ۧ + … + 𝐶𝑛ψ𝑛 𝐶𝑛ψ𝑛 ۧ
𝐶12 E1 + 𝐶22 E2 + ⋯ + 𝐶𝑛2 En If 𝞥 is normalized, the denominator σ𝑛𝑖=1 𝐶𝑖2 = 1
=
𝐶12 + 𝐶22 + … + 𝐶𝑛2
𝑛 Subtracting ℰ from the true ground state energy Ei we have
ℰ = ෍ 𝐶𝑖2 𝐸𝑖 ℰ − E1 = σ𝑛𝑖=1 𝐶𝑖2 𝐸𝑖 − E1
𝑖=1 = σ𝑛𝑖=1 𝐶𝑖2 𝐸𝑖 − E1 σ𝑛𝑖=1 𝐶𝑖2 = σ𝑛𝑖=1 𝐶𝑖2 𝐸𝑖 − E1 ≥ 0
Best case Scenario, 𝐶1 = 1 𝑎𝑛𝑑 𝐶2 , 𝐶3 , … 𝐶𝑛 all equal to zero
 Linear Variation
The trial wave function 𝞥 that is expressed as a linear combination (superposition) of n linearly
independent functions ψ1 , ψ2 , ψ3 , ψ4 …. ψ𝑛 , i.e., 𝑛

𝞥 = 𝐶1ψ1 + 𝐶2ψ2 + 𝐶3ψ3 + ⋯ 𝐶𝑛ψ𝑛 = ෍ 𝐶𝑖ψ𝑖

𝑖=1
The ψ𝑖 are known functions and the scaling coefficients 𝐶𝑖 are variational 𝞥𝐻𝞥
parameters (often complex) to be determined by minimizing the energy ℰ=
𝞥𝞥
expectation value associated with 𝞥. Substituting 𝞥 We have
σ𝑛𝑖=1 𝐶𝑖ψ𝑖 𝐻 σ𝑛𝑖=1 𝐶𝑖ψ𝑖 σ𝑛𝑖=1 𝐶𝑖2 ψ𝑖 𝐻 ψ𝑖 + σ𝑛𝑖≠𝑗 𝐶𝑖 𝐶𝑗 ψ𝑖 𝐻 ψ𝑗 σ𝑛𝑖=1 𝐶𝑖2 𝐻𝑖𝑖 + σ𝑛𝑖<𝑗 2𝐶𝑖 𝐶𝑗𝐻𝑖𝑗 ψ𝑖 𝐻 ψ𝑗 = 𝐻𝑖𝑗
ℰ= = = 𝑛
σ𝑛𝑖=1 𝐶𝑖ψ𝑖 σ𝑛𝑖=1 𝐶𝑖ψ𝑖 σ𝑛𝑖=1 𝐶𝑖2 ψ𝑖 ψ𝑖 + σ𝑛𝑖≠𝑗 𝐶𝑖 𝐶𝑗 ψ𝑖 ψ𝑗 σ𝑖=1 𝐶𝑖2 𝑆𝑖𝑖 + σ𝑛𝑖<𝑗 2𝐶𝑖 𝐶𝑗𝑆𝑖𝑗 ψ𝑖 ψ𝑗 ൿ = Sij
δℰ
For minimizing ℰ, we have to find the optimum value 𝐶𝑖 by setting to zero for all i.
δ𝐶𝑖
δℰ u δu δv 𝑢 1 δu
= = − ∗ =0 = 2𝐶𝑖𝐻𝑖𝑖 + σ𝑛𝑖<𝑗 2𝐶𝑗 𝐻𝑖𝑗
δ𝐶𝑖
δ𝐶𝑖 v δ𝐶𝑖 δ𝐶𝑖 𝑣 𝑣
1 δu δv 𝑢 δv
Since cannot be zero we have − =0 δ𝐶𝑖
= 2𝐶𝑖𝑆𝑖𝑖 + σ𝑛𝑖<𝑗 2𝐶𝑗 𝑆𝑖𝑗
𝑣 δ𝐶𝑖 δ𝐶𝑖 𝑣
𝑢
𝑛 =ℰ
δℰ 𝑣
Substituting we have = ෍ 𝐶𝑗 (𝐻𝑖𝑗 −ℰ𝑆𝑖𝑗 ) = 0
δ𝐶𝑖
𝑗=1
 Secular Determinant
δℰ
Minimization of the 𝐶𝑖 in the trial wave functionσ𝑛
𝑖=1 𝐶𝑖 ψ𝑖 leads to = σ𝑛𝑗=1 𝐶𝑗 (𝐻𝑖𝑗 −ℰ𝑆𝑖𝑗 ) = 0
δ𝐶𝑖
For n variation coefficients, we have n such equations which can be expanded as
𝐶1(𝐻11 −ℰ𝑆11 ) + 𝐶2(𝐻12 −ℰ𝑆12 ) + ⋯ + 𝐶𝑛(𝐻1𝑛 −ℰ𝑆1𝑛 ) = 0
ψ𝑖 𝐻 ψ𝑗 = 𝐻𝑖𝑗
𝐶1(𝐻21 −ℰ𝑆21 ) + 𝐶2(𝐻22 −ℰ𝑆12 ) + ⋯ + 𝐶𝑛(𝐻2𝑛 −ℰ𝑆2𝑛 ) = 0
ψ𝑖 ψ𝑗 ൿ = Sij
⁞
𝐶1(𝐻𝑛1 −ℰ𝑆𝑛1 ) + 𝐶2(𝐻𝑛2 −ℰ𝑆𝑛2 ) + ⋯ + 𝐶𝑛(𝐻𝑛𝑛 −ℰ𝑆𝑛𝑛 ) = 0
Since there are linear equations with n unknowns, they can be solved using matrix operationalization.
𝐻11 − ℰ𝑆11 𝐻12 − ℰ𝑆12 ⋯ 𝐻1𝑛 − ℰ𝑆1𝑛 The trivial solution 𝐶1, 𝐶2, … 𝐶𝑛 all zero is ignored.
𝐶1
𝐻21 −ℰ𝑆21 ⋮
⋱ ⋮ =0
⋮ 𝐶𝑛
Necessary and Sufficient Condition
𝐻𝑛1 − ℰ𝑆𝑛1 𝐻𝑛2 − ℰ𝑆𝑛2 ⋯ 𝐻𝑛𝑛 − ℰ𝑆𝑛𝑛 Non-trivial solution only if the determinant vanish.
Hence
𝐻11 − ℰ𝑆11 𝐻12 − ℰ𝑆12 ⋯ 𝐻1𝑛 − ℰ𝑆1𝑛 Leads to the characteristic polynomial in ℰ of order n
𝐻21 −ℰ𝑆21 n real roots to ℰ , corresponding to n Eigen values
⋱ ⋮ =0
⋮
𝐻𝑛1 − ℰ𝑆𝑛1 𝐻𝑛2 − ℰ𝑆𝑛2 ⋯ 𝐻𝑛𝑛 − ℰ𝑆𝑛𝑛 Not only ℰ1 ≥ 𝐸1 , ℰ𝑛 ≥ 𝐸𝑛 for any nth excited state.
Each ℰ n, solve for the coefficients for ψ𝑖 which gives the ratio between coefficients 𝐶𝑖 in 𝞥n
Absolute values of 𝐶𝑖 obtained by Normalization. Degenerate 𝞥n requires Schmidt orthogonalization
 Illustration of Linear Variation: Hydrogen in z-directed Electric Field
in electric field F, Hydrogen’s centre of electron density moves away from nuclei
The redistribution of electron density causes polarization.
This develops dipole moment μ ∝ F or μ = α F, where α – polarizability
Skewed electron density can be approximated by a Linear trial wave function
𝞥 = 𝐶11𝑠 + 𝐶22𝑝𝑧
1
The Hamiltonian operator is 𝐻 = − 2 ∇2 − 𝑟 − eFz = HH + Fr cos 
1
1
The Secular Determinant is 1
𝐻11 = < 1𝑠| 𝐻 1𝑠 > = < 1𝑠 𝐻𝐻 |1𝑠 > +< 1𝑠|Fr cos |1𝑠 >= − + 0
2
𝐻11 − ℰ𝑆11 𝐻12 − ℰ𝑆12 (r cos  is antisymmetric while 1s*1s is symmetric)
=0
𝐻21 − ℰ𝑆21 𝐻22 − ℰ𝑆22 1
15
𝐻22 = < 2𝑝𝑧 | 𝐻 2𝑝𝑧 > = < 2𝑝𝑧 𝐻𝐻 |2𝑝𝑧 > +< 2𝑝𝑧 |𝐹r cos |2𝑝𝑧 > = − 8 + 0
1 22
− −ℰ − 𝐹 𝐻12 = 𝐻21 = < 1𝑠| 𝐻 2𝑝𝑧 > = < 1𝑠 𝐻𝐻 |2𝑝𝑧 > + < 1𝑠|Fr cos |2𝑝𝑧 > = 0 + ?
2 35 =0
15 1 −𝑟 1 𝑟
−2
22 1 < 1𝑠|r cos |2𝑝𝑧 >= −𝐹 ම 𝑒 (𝑟 𝑐𝑜𝑠 ) 𝑟𝑒 cos r2 𝑑𝑟 𝑠𝑖𝑛 𝑑 𝑑
− 5 𝐹 − −ℰ 𝜋 32𝜋 15
3 8 1 ∞ 3𝑟
−2 𝜋 𝐹 4! 2 22
= −𝐹 𝜋 2𝜋 ‫׬‬0 𝑟 4 𝑒 𝑑𝑟 ‫׬‬0 cos2  𝑠𝑖𝑛 𝑑 = − 2 3 5
= − 𝐹
2
5 15
F2 1 32 2 3 35
ℰ + ℰ−2 + =0 2
8 310 16 𝑆11 =< 1𝑠| 1𝑠 > = 1 𝐷𝑢𝑒 𝑡𝑜 𝑡ℎ𝑒 𝑜𝑟𝑡ℎ𝑜𝑛𝑜𝑟𝑚𝑎𝑙 𝑏𝑎𝑠𝑖𝑠, 𝑠𝑒𝑣𝑒𝑟𝑎𝑙
5 9 217 𝐹 2 𝑆12 = < 1𝑠| 2𝑝𝑧 > = 0 = 𝑆21 𝑖𝑛𝑡𝑒𝑟𝑔𝑟𝑎𝑙𝑠 𝑎𝑟𝑒 𝑒𝑎𝑠𝑦 𝑡𝑜 𝑒𝑣𝑎𝑙𝑢𝑎𝑡𝑒.
− 8 ± 64 + 10
3
ℰ= 𝑆22 =< 2𝑝𝑧 |2𝑝𝑧 > = 1
2
 Polarized Hydrogen: wavefunctions
5 9 217 𝐹 2 When the applied field strength F = 0, we get back the original energies -1/2 a.u. and -1/8 a.u.
− 8 ± 64 + 10
3
ℰ= For the F= 0.1a.u, the values are ℰ= -0.51425 (minimum in ℰ vs C1 ) and ℰ = -0.1107 (maximum)
2
𝐶1(𝐻11 −𝐸) + 𝐶2𝐻12 = 0
𝐶1𝐻12 + 𝐶2(𝐻22 −𝐸) = 0
For ℰ= -0.51425, we have
0.01425𝐶1 − 0.07449𝐶2 = 0
𝐶1 = 5.2275𝐶2
−0.07449𝐶1 + 0.38925𝐶2 = 0
Using normalization (σ𝑛𝑖=1 𝐶𝑖2 = 1) we have
5.2275 𝐶22 + 𝐶22 = 1 𝐶2 = ±0.18789
Taking positive 𝐶2 we have 𝐶1 = 0.98219

Variation of ℰ with 𝐹 Variation of C1 with ℰ

For ℰ= -0.51425, we have
𝞥+ = 0.982191𝑠 + 0.18789 2𝑝𝑧
For ℰ= -0.1107, we have
𝞥− = 0.982192𝑝𝑧 − 0.187891𝑠
 Illustration for Non-linear Variation: Hydrogen atom
Trial wave function for Hydrogen atom (a) spherical symmetry (b) exponential decay in r.
෡ = − ∇2 − 1 1
The Hamiltonian operator is 𝐻 2 𝑟
𝜁 3 −𝜁𝑟
Normalized 𝞥𝑡𝑟𝑖𝑎𝑙 = 𝑒 11 𝜕 2
𝜕 1 𝜕 𝜕 1 𝝏𝟐 1
𝜋 =− 2 𝑟 + 𝑠𝑖𝑛θ + 2 −
2 𝑟 𝜕𝑟 𝜕𝑟 𝑠𝑖𝑛 θ 𝜕θ 𝜕θ sin θ 𝝏𝝋𝟐 𝑟

𝜁 3 −𝜁𝑟 1 1 𝜕 2 𝜕 1 𝜕 𝜕 1 𝝏𝟐 1 𝜁 3 −𝜁𝑟
ℰ= 𝑒 − 𝑟 + 𝑠𝑖𝑛θ + −
𝜋 2 𝑟 2 𝜕𝑟 𝜕𝑟 𝑠𝑖𝑛 θ 𝜕θ 𝜕θ sin2 θ 𝝏𝝋𝟐 𝑟1 𝜋𝑒
𝜁3 −𝜁𝑟
1 1 𝜕 2
𝜕 1 𝜕 𝜕 1 𝝏𝟐
1 −𝜁𝑟 2
= ම𝑒 (− 2 𝑟 + 𝑠𝑖𝑛θ + 2 − )𝑒 r 𝑑𝑟 𝑠𝑖𝑛 𝑑 𝑑
𝜋 2 𝑟 𝜕𝑟 𝜕𝑟 𝑠𝑖𝑛 θ 𝜕θ 𝜕θ sin θ 𝝏𝝋𝟐 𝑟1
∞
𝜁3 11 𝜕 𝜕 1 −𝜁𝑟 2 𝑂𝑝𝑡𝑖𝑚𝑖𝑧𝑖𝑛𝑔 𝜁
= න 𝑒 −𝜁𝑟 − 𝑟 2 − 𝑒 r 𝑑𝑟 ඵ 𝑠𝑖𝑛 𝑑 𝑑
𝜋 2 𝑟 2 𝜕𝑟 𝜕𝑟 𝑟1
0
∞
𝜁2
𝜁3 𝜁 − 1 𝜁 2 −𝜁𝑟 2 ℰ= −𝜁
−𝜁𝑟 2
= 4𝜋 න 𝑒 − 𝑒 r 𝑑𝑟
𝜋 r 2 𝜕𝐸
0 =𝜁−1=0
∞ ∞ 𝑛! 𝜕𝜁
𝜁 − 1 𝜁2 2 ‫׬‬0 x n 𝑒 −𝑎𝑥 𝑑𝑥 = 𝑎𝑛+1
= 4𝜁 3 න 𝑒 −2𝜁𝑟 − r 𝑑𝑟 𝜁=1
r 2
0
∞ ∞
3 −2𝜁𝑟
𝜁2 𝜁2
= 4𝜁 (𝜁 − 1) න 𝑒 r𝑑𝑟 − න 𝑒 −2𝜁𝑟 r 2 𝑑𝑟 ℰ= −𝜁 Trial wave function leads to exact solution!
2 2
0 0
Time Independent Perturbational Method : The Recipe (Non-Degenerate)
Step 1: Dissect the Hamiltonian H into a known 𝐻0 (zeroth Order) and an unknown 𝐻′ (Perturbation term)
෡=𝐻
𝐻 ෡ 0 + λ𝐻′ ෡ 𝑛 = E𝑛ψ𝑛
𝐻ψ ෡ 0 ψ0𝑛 = 𝐸𝑛0 ψ0𝑛
𝐻 ෡ 0 + λ𝐻′ )ψ𝑛 = E𝑛ψ𝑛
(𝐻
෡ ′ is small, Expand Energies En and wavefunctions ψn as Taylor series
Step 2: Assuming the perturbation 𝐻
E𝑛 = 𝐸𝑛0 + λ𝐸𝑛𝐼 + λ2 𝐸𝑛𝐼𝐼 + … … ψ𝑛 = ψ0𝑛 + λψ𝐼𝑛 + λ2 ψ𝐼𝐼
𝑛 + ……

Where the handle λ values are 0 < λ < 1. Substituting We have

෡ 0 + λ𝐻′ ( ψ0𝑛 + λψ𝐼𝑛 + λ2 ψ𝐼𝐼
𝐻 0 𝐼 2 𝐼𝐼 0 𝐼 2 𝐼𝐼
𝑛 + … ) = 𝐸𝑛 + λ𝐸𝑛 + λ 𝐸𝑛 + ⋯ ( ψ𝑛 + λψ𝑛 + λ ψ𝑛 + ⋯ )

෡→𝐻
When λ → 0, 𝐻 ෡ 0 . Collecting the coefficients of similar powers of λ , We have
෡ 0 ψ0𝑛 − 𝐸𝑛0 ψ0𝑛 + λ 𝐻′ ψ0𝑛 + 𝐻
𝐻 ෡ 0 ψ1𝑛 − 𝐸𝑛0 ψ𝐼𝑛 − 𝐸𝑛1 ψ0𝑛 + λ2 𝐻
෡ 0 ψ𝐼𝐼 ′ 𝐼 0 𝐼𝐼 1 𝐼 2 0
𝑛 + 𝐻 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 … = 0

This equation holds only if the coefficients of every individual powers of λ became zero
෡ 0 ψ0𝑛 + 𝐸𝑛0 ψ0𝑛 = 0
λ0 ∶ 𝐻
෡ 0 ψ1𝑛 − 𝐸𝑛0 ψ𝐼𝑛 − 𝐸𝑛1 ψ0𝑛 = 0
λ1 ∶ 𝐻′ ψ0𝑛 + 𝐻
෡ 0 ψ𝐼𝐼
λ2 ∶ 𝐻 ′ 𝐼 0 𝐼𝐼 1 𝐼 2 0
𝑛 + 𝐻 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 = 0

Step 3: Evaluate the Correction to the Energies and Wavefunctions up to the required order.
 Time Independent Perturbational Method : Ist Order Correction
The First order equation is λ1 ∶ ෡ ′ ψ0𝑛 + 𝐻
𝐻 ෡ 0 ψ𝐼𝑛 − 𝐸𝑛0 ψ𝐼𝑛 − 𝐸𝑛1 ψ0𝑛 = 0

Energy 𝐸𝑛𝐼 : Left Multiplying by ψ∗0𝑙 and integrating over the domain we have
෡ ′ ψ0𝑛
ψ0𝑙 𝐻 ෡ 0 ψ1𝑛 − 𝐸𝑛0 ψ0𝑙 ψ𝐼𝑛 ൿ − 𝐸𝑛𝐼 ψ0𝑙 ψ0𝑛 = 0
+ ψ0𝑙 𝐻

When l=n, ෡ ′ ψ0𝑛

ψ0𝑛 𝐻 + ψ1𝑛 𝐻0 ψ0𝑛 − 𝐸𝑛0 ψ0𝑛 ψ𝐼𝑛 ۧ − 𝐸𝑛𝐼 ψ0𝑛 ψ0𝑛 = 0 ෡ ′ ψ0𝑛
𝐸𝑛𝐼 = ψ0𝑛 𝐻

Wave Function ψ𝐼𝑛 : Expand ψ𝐼𝑛 as linear combinations of orthonormal zeroth order functions
ψ𝐼𝑛 = 𝐶1𝑛 ψ10 + 𝐶2𝑛 ψ02 + 𝐶3𝑛 ψ03 + 𝐶4𝑛 ψ04 + ⋯ = ෍ 𝐶𝑚𝑛ψ0𝑚 We need to find the values of 𝐶𝑚𝑛
𝑚
Substituting in the first order equation we have
෡ ′ ψ0𝑛 + 𝐻
𝐻 ෡ 0 σ𝑚 𝐶𝑚𝑛ψ0𝑚 − 𝐸𝑛0 σ𝑚 𝐶𝑚𝑛ψ0𝑚 − 𝐸𝑛1 ψ0𝑛 = 0 ෡ ′ ψ0𝑛
ψ0𝑚 𝐻
𝐶𝑚𝑛 =
𝐸𝑛0 − 𝐸𝑚0
Left Multiplying by ψ∗0𝑙 and integrating over the domain we have
෡ 0 σ𝑚 𝐶𝑚𝑛ψ0𝑖 − 𝐸𝑛0 ψ0𝑙 σ𝑚 𝐶𝑚𝑛ψ0𝑚 − 𝐸𝑛𝐼 ψ0𝑙 ψ0𝑛 = 0
෡ ′ ψ0𝑛 + ψ0𝑙 𝐻
ψ0𝑙 𝐻
෡ ′ ψ0𝑛
ψ0𝑚 𝐻
෡ ′ ψ0𝑛 + 𝐶𝑙𝑛𝐸𝑙0 ψ0𝑙 ψ0𝑙 − 𝐶𝑙𝑛𝐸𝑛0 ψ0𝑙 ψ0𝑙 − 𝐸𝑛𝐼 ψ0𝑙 ψ0𝑛 = 0 ψ𝐼𝑛 = ෍ ψ0𝑚
When l≠n, ψ0𝑙 𝐻 𝐸𝑛0 − 𝐸𝑚0
𝑚≠𝑛

෡ ′ ψ0𝑛 + 𝐶𝑙𝑛 𝐸𝑙0 − 𝐶𝑙𝑛𝐸𝑛0 = 0

ψ0𝑙 𝐻
 IInd Order Correction to Energy 𝐸𝑛𝐼𝐼
The Second order equation is λ2 ∶ ෡ 0 ψ𝐼𝐼
𝐻 ′ 𝐼 0 𝐼𝐼 𝐼 𝐼 𝐼𝐼 0
𝑛 + 𝐻 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 = 0

Left Multiplying by ψ∗0𝑙 and integrating over the domain we have

෡ 0 ψ𝐼𝐼
ψ0𝑙 𝐻 0 ෡′ 𝐼
𝑛 + ψ𝑙 𝐻 ψ𝑛 − 𝐸𝑛0 ψ0𝑙 ψ𝐼𝐼 𝐼 0 𝐼 𝐼𝐼 0 0
𝑛 ൿ − 𝐸𝑛 ψ𝑙 ψ𝑛 − 𝐸𝑛 ψ𝑙 ψ𝑛 = 0

When l = n, ෡ 0 ψ𝐼𝐼
ψ0𝑛 𝐻 𝑛 + ψ0𝑛 𝐻′ ψ𝐼𝑛 − 𝐸𝑛0 ψ0𝑛 ψ𝐼𝐼 𝐼 0 𝐼 𝐼𝐼 0 0
𝑛 ۧ − 𝐸𝑛 ψ𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 ψ𝑛 = 0

ψ𝐼𝐼 ෡0 0
𝑛 𝐻 ψ𝑛 + ψ0𝑛 𝐻′ ψ𝐼𝑛 − 𝐸𝑛0 ψ0𝑛 ψ𝐼𝐼 𝐼 0 𝐼 𝐼𝐼
𝑛 ۧ − 𝐸𝑛 ψ𝑛 ψ𝑛 − 𝐸𝑛 = 0

Substituting the value of ψ𝐼𝑛 from the first order equation we have
෡ ′ ψ0𝑛
ψ0𝑚 𝐻 ෡ ′ ψ0𝑛
ψ0𝑙 𝐻
ψ0𝑛 𝐻′ ෍ ψ0𝑚 − 𝐸𝑛𝐼 ψ0𝑛 σ𝑚≠𝑛 0 ψ0𝑚 − 𝐸𝑛𝐼𝐼 = 0
𝐸𝑚0
− 𝐸𝑛0 𝐸𝑚 − 𝐸𝑛0
𝑚≠𝑛

෡ ′ ψ0𝑚 ψ0𝑚 𝐻
ψ0𝑛 𝐻 ෡ ′ ψ0𝑛
෍ 0 − 𝐸𝑛𝐼𝐼 = 0
𝐸𝑚 − 𝐸𝑛0
𝑚≠𝑛
෡ ′ ψ0𝑚 ψ0𝑚 𝐻
ψ0𝑛 𝐻 ෡ ′ ψ0𝑛 ෡𝑚𝑛
𝐻 ′ 𝐻෡𝑛𝑚
′
𝐸𝑛𝐼𝐼 = ෍ = ෍
𝐸𝑛𝐼𝐼 turns out to be a series, truncated at some point. 𝐸𝑛0 − 𝐸𝑚
0
𝐸𝑛0 − 𝐸𝑚0
𝑚≠𝑛 𝑚≠𝑛
 IInd Order Correction to Wave functions
The Second order equation is λ2 ∶ ෡ 0 ψ𝐼𝐼
𝐻 ′ 𝐼 0 𝐼𝐼 𝐼 𝐼 𝐼𝐼 0
𝑛 + 𝐻 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 = 0
Left Multiplying by ψ∗0𝑙 and integrating over the domain we have
0 ෡ 0 𝐼𝐼 0 ෡′ 𝐼 0 0 𝐼𝐼 𝐼 0 𝐼 𝐼𝐼 0 0 need to be determined
ψ𝑙 𝐻 ψ𝑛 + ψ𝑙 𝐻 ψ𝑛 − 𝐸𝑛 ψ𝑙 ψ𝑛 ൿ − 𝐸𝑛 ψ𝑙 ψ𝑛 − 𝐸𝑛 ψ𝑙 ψ𝑛 = 0
Expand ψ𝐼𝐼𝑛 as linear combinations of orthonormal zeroth order functions ψ𝐼𝐼
𝑛 = ෍ 𝐶 𝑚𝑛 ψ0
𝑚

Substituting ψ𝐼𝐼
𝑛 in the second order equation we have 𝑚

෡ ′ ψ𝐼𝑛 − 𝐸𝑛0 ψ0𝑙 σ𝑚 𝐶𝑚𝑛ψ0𝑚 − 𝐸𝑛𝐼 ψ0𝑙 ψ𝐼𝑛 − 𝐸𝑛𝐼𝐼 ψ0𝑙 ψ0𝑛 = 0
ψ0𝑙 𝐻0 σ𝑚 𝐶𝑚𝑛ψ0𝑚 + ψ0𝑙 𝐻

When l≠n, ෡ ′ ψ𝐼𝑛

𝐶𝑙𝑛 ψ0𝑙 𝐻0 ψ0𝑙 + ψ0𝑙 𝐻 − 𝐸𝑛0 𝐶𝑙𝑛 ψ0𝑙 ψ0𝑙 − 𝐸𝑛𝐼 ψ0𝑙 ψ𝐼𝑛 − 𝐸𝑛𝐼𝐼 ψ0𝑙 ψ0𝑛 = 0
෡ ′ ψ𝐼𝑛
𝐶𝑙𝑛 (𝐸𝑙0 −𝐸𝑛0 ) + ψ0𝑙 𝐻 − 𝐸𝑛𝐼 ψ0𝑙 ψ𝐼𝑛 = 0
෡ ′ ψ0𝑛
ψ0𝑘 𝐻 ෡ ′ ψ0𝑛
ψ0𝑚 𝐻
Substituting ψ𝐼𝑛 & 𝐸𝑛𝐼 𝐶𝑙𝑛 (𝐸𝑙0 −𝐸𝑛0 ) + ψ0𝑙 𝐻′ ෍ ψ0𝑘 − ψ0𝑛 ෡ ′ ψ0𝑛 ψ0𝑙 σ𝑚≠𝑛
𝐻 ψ0𝑚 = 0
𝐸𝑛0 − 𝐸𝑘0 𝐸𝑛0 − 𝐸𝑚0
𝑘≠𝑛
෡ ′ ψ0𝑘 ψ0𝑘 𝐻
ψ0𝑙 𝐻 ෡ ′ ψ0𝑛 ψ0𝑛 𝐻෡′ ψ0𝑛 ψ0𝑙 𝐻 ෡ ′ ψ0𝑛
𝐶𝑙𝑛 (𝐸𝑙0 −𝐸𝑛0 ) + ෍ − =0
𝐸𝑛0 − 𝐸𝑘0 𝐸𝑛0 − 𝐸𝑙0
𝑘≠𝑛
෡𝑙𝑘
𝐻 ′ ෡′
𝐻𝑘𝑛 ෡𝑛𝑛
𝐻 ෡𝑚𝑛
′ 𝐻 ′
𝐶𝑙𝑛 (𝐸𝑙0 −𝐸𝑛0 ) +෍ 0 − 0 =0 ෡𝑚𝑘
𝐻 ′ ෡′
𝐻𝑘𝑛 ෡𝑛𝑛
𝐻 ෡𝑚𝑛
′ 𝐻 ′
𝐸𝑛 − 𝐸𝑘0 𝐸𝑛 − 𝐸𝑙0 ψ𝐼𝐼
𝑛 = ෍ ෍ − ψ0𝑚
𝑘≠𝑛 𝐸𝑛0 − 𝐸𝑘0 (𝐸𝑛0 −𝐸𝑚
0
) 𝐸𝑛0 − 𝐸𝑚 0 2
𝑚≠𝑛 𝑘≠𝑛
෡𝑙𝑘
𝐻 ′ ෡′
𝐻𝑘𝑛 𝐻෡𝑛𝑛
′ ෡′
𝐻𝑙𝑛
𝐶𝑙𝑛 = ෍ −
𝐸𝑛0 − 𝐸𝑘0 (𝐸𝑛0 −𝐸𝑙0 ) 𝐸𝑛0 − 𝐸𝑙0 2
𝐸𝑛𝐼𝐼 turns out to be a double summation series, truncated at some point.
𝑘≠𝑛
 (Time independent) Degenerate Perturbation
෡ 0 ψ0𝑛 = 𝐸𝑛0 ψ0𝑛 , where 𝐸𝑛0 is d fold degenerate i.e., 𝐸10 = 𝐸20 = 𝐸30 = ⋯ 𝐸𝑑0
If the unperturbed SE 𝐻
෡=𝐻
The perturbed system 𝐻 ෡ 0 + λ𝐻′ ෡ 𝑛 = E𝑛ψ𝑛
𝐻ψ ෡ 0 + λ𝐻′ )ψ𝑛 = E𝑛ψ𝑛
(𝐻
When λ → 0, E𝑛 → 𝐸10 but ψ𝑛 → ψ0𝑛 ?,
but ψ𝑛 → 0𝑛 = σ𝑑 𝐶𝑑 ψ0𝑑 where 𝐶𝑑 are different for each ψ𝑛

The series expansions are E𝑛 = 𝐸𝑑0 + λ𝐸𝑛𝐼 + λ2 𝐸𝑛𝐼𝐼 + …

ψ𝑛 = 0𝑛 + λψ𝐼𝑛 + λ2 ψ𝐼𝐼 0 𝐼 2 𝐼𝐼
𝑛 + … = ෍ 𝐶𝑖 ψ𝑖 + λψ𝑛 + λ ψ𝑛 + …
𝑑
Substituting and Collecting the coefficients of similar powers of λ , We have

෡ 0 0𝑛 + 𝐸𝑛0 0𝑛 = 0

λ0 ∶ 𝐻

λ1 ∶ 𝐻′ 0𝑛 + 𝐻
෡ 0 ψ1𝑛 − 𝐸𝑛0 ψ𝐼𝑛 − 𝐸𝑛1 0𝑛 = 0

𝑛 + 𝐻 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 ψ𝑛 − 𝐸𝑛 𝑛 = 0
෡ 0 ψ𝐼𝐼
λ2 ∶ 𝐻 ′ 𝐼 0 𝐼𝐼 1 𝐼 2 0
 (First Order) Degenerate Perturbation
λ1 ∶ 𝐻′ 0𝑛 + 𝐻
෡ 0 ψ1𝑛 − 𝐸𝑛0 ψ𝐼𝑛 − 𝐸𝑛1 0𝑛 = 0
First Order Energy
Left Multiplying by ψ∗0𝑙 and integrating over the domain we have
෡ ′ 0𝑛
ψ0𝑙 𝐻 ෡ 0 ψ1𝑛 − 𝐸𝑛0 ψ0𝑙 ψ𝐼𝑛 ൿ − 𝐸𝑛𝐼 ψ0𝑙 0𝑛 = 0
+ ψ0𝑙 𝐻

When l=n, ෡ ′ 0𝑛

ψ0𝑛 𝐻 + ψ1𝑛 𝐻0 0𝑛 − 𝐸𝑛0 0𝑛 ψ𝐼𝑛 ۧ − 𝐸𝑛𝐼 ψ0𝑛 0𝑛 = 0

෡ ′ σ𝒅𝒊=𝟏 𝑪𝒊ψ𝟎𝒊 − 𝑬𝑰𝒏 ψ𝟎𝒏 σ𝒅𝒊=𝟏 𝑪𝒊ψ𝟎𝒊 = 0

ψ𝟎𝒏 𝑯
෡ 11 −𝑬𝑰𝒏 ) + 𝑪𝟐 𝑯
𝑪𝟏 (𝑯 ෡ 12 + 𝑪𝟑 𝑯
෡ 13 + ⋯ 𝑪𝒅 𝑯
෡ 1d = 0
σ𝒅𝒊=𝟏 𝑪𝒊 ψ𝟎𝒏 𝑯
෡ ′ ψ𝟎𝒊 − 𝑬𝑰𝒏 σ𝒅𝒊=𝟏 𝑪𝒊 ψ𝟎𝒏 ψ𝟎𝒊 = 0
෡ 21 + 𝑪𝟐 (𝑯
𝑪𝟏 𝑯 ෡ 22 −𝑬𝑰𝒏 ) + 𝑪𝟑 𝑯
෡ 23 + ⋯ 𝑪𝒅 𝑯
෡ 2𝑑 = 0

σ𝒅𝒊=𝟏 𝑪𝒊 ( ψ𝟎𝒏 𝑯
෡ ′ ψ𝟎𝒊 − 𝑬𝑰𝒏 δ ) = 0
𝒏𝒊
෡ 31 + 𝑪𝟐 𝑯
𝑪𝟏 𝑯 ෡ 32 + 𝑪𝟑 (𝑯
෡ 23 −𝑬𝑰𝒏 ) + ⋯ 𝑪𝒅 𝑯
෡ 3𝑑 = 0
: : :
෡ ′ ψ𝟎𝒊 = 𝑯
Substituting ψ𝟎𝒏 𝑯 ෡ ni we have : : :
෡ d1 + 𝑪𝟐 𝑯
𝑪𝟏 𝑯 ෡ 𝑑2 + 𝑪𝟑 𝑯
෡ 23 + ⋯ 𝑪𝒅 (𝑯
෡ 3𝑑 −𝑬𝑰𝒏 ) = 0
σ𝒅𝒊=𝟏 𝑪𝒊 (𝑯
෡ ni −𝑬𝑰𝒏 δ ) = 0
𝒏𝒊
 Degenerate Perturbation Leads to Determinental Solutions

𝐻11 − 𝑬𝑰𝒏 𝐻12 ⋯ 𝐻1𝑛

𝐶1 The trivial solution 𝐶1, 𝐶2, … 𝐶𝑛 all zero is ignored.
𝐻21 ⋮
⋱ ⋮ =0
⋮
𝐻𝑛1 𝐻𝑛2 ⋯ 𝐻𝑛𝑛 − 𝑬𝑰𝒏
𝐶𝑛 Necessary and Sufficient Condition
Non-trivial solution only if the determinant vanish.

𝐻11 − 𝑬𝑰𝒏 𝐻12 ⋯ 𝐻1𝑛 Expansion gives a polynomial of degree d

𝐻21 Root of the polynomial determines the splitting of degeneracy.
⋱ ⋮ =0
⋮
𝐻𝑛1 𝐻𝑛2 ⋯ 𝐻𝑛𝑛 − 𝑬𝑰𝒏 Substituting the roots back in the determinant and using
normalization condition allows Evaluation of the coefficients (Ci).

If all the solutions are identical, this implies the perturbation is not affecting the degeneracy
Otherwise, this d fold degeneracy is split into various energies.
If the determinant is diagonal .

it implies that the assumed first order wavefunctions are correct.

 Helium Atom : Perturbation theory
Assuming the nucleus is fixed we have 1 r12
2
1 2 1 2 𝑧 𝑧 1 1
𝐻 = − ∇1 − ∇2 − − + =𝐻 1 +𝐻 2 + r1
2 2 𝑟1 𝑟2 𝑟12 𝑟12 r2
෡0
𝐻 ෡′
𝐻

෡=𝐻
𝐻 ෡ 0 + λ𝐻′ ෡ 𝑛 = E𝑛ψ𝑛
𝐻ψ ෡ 0 ψ0𝑛 = 𝐸𝑛0 ψ0𝑛
𝐻 ෡ 0 + λ𝐻′ )ψ𝑛 = E𝑛ψ𝑛
(𝐻

𝑍2 𝑍2
ψ10 = 1𝑠 1 ∗ 1𝑠(2) E10 = + = 𝑍 2 (in atomic units) (-108.83 eV)
2 2

෡ ′ ψ0𝑛 = 1𝑠 1 ∗ 1𝑠(2) 1
The First order correction to the Energy 𝐸𝑛𝐼 = ψ0𝑛 𝐻 1𝑠 1 ∗ 1𝑠(2)
𝑟12
2𝜋 𝜋 ∞
𝑧 3 −𝑧𝑟 𝑧 3 −𝑧𝑟 1 𝑧 3 −𝑧𝑟 𝑧 3 −𝑧𝑟 2
𝐸𝑛𝐼 = න නන 𝑒 1 𝑒 2 𝑒 1 𝑒 2 𝑟1 𝑑𝑟1 𝑟22 𝑑𝑟2 sin 𝜃1 𝑑𝜃1 sin 𝜃2 𝑑𝜃2 𝑑𝜑2 𝑑𝜑2
𝜋 𝜋 𝑟12 𝜋 𝜋
0 0 0

5 Exp
= 𝑧 (in atomic units) (34.01 eV)
8
5
Energy after first order correction = E10 + E1𝐼 = 𝑍 2 + 8 𝑧 (-74.82 eV) (5% error) E

Experimental Energy -79.0 eV E1𝐼𝐼 = −4.29 𝑒𝑉 E1𝐼𝐼 = 0.12𝑒𝑉

 Helium Atom : Variation theory
𝞥𝐻𝞥 r12
𝑉𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦 ℰ= ≥ 𝐸𝑔 1
𝞥𝞥 2
Trial Wave Function ψtrial? r1
+ r2
Hydrogenic wave functions as trial wave function 𝞥𝑔0 = 1𝑠 1 ∗ 1𝑠(2) (from He )
1 1
𝞥𝑔0 𝐻 𝞥𝑔0 = 𝞥𝑔0 𝐻 1 + 𝐻 2 + 𝑟 𝞥𝑔0 = 𝞥𝑔0 𝐻 1 𝞥𝑔0 + 𝞥𝑔0 𝐻 2 𝞥𝑔0 + 𝞥𝑔0 𝑟 𝞥𝑔0
12 12
5
= 2𝐸1𝑠 + 𝑧 = 𝐸𝑔0 +𝐸𝑔𝐼 (Same as First order correction in perturbation theory)
8
Alternate ψtrial?
Chemical Intuition: Helium 1s orbitals should be more diffuse than He+
The power of the exponent should be declared variational parameter as it controls diffusion.

𝑧 3 −𝑧𝑟 (𝑧 = 𝑛𝑢𝑐𝑙𝑒𝑎𝑟 𝑐ℎ𝑎𝑟𝑔𝑒)

Ψg Hydrogenic = 𝑒 1
𝜋
Declare Nuclear charge as a variational parameter leading to non-linear variation (Effective Nuclear Charge)

3 −𝑟 ( = 𝑣𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟)

Ψg modified = 𝑒 1
𝜋
 Helium Atom : Variation theory
3 −𝑟 3 −𝑟 𝞥𝐻𝞥
Trial wave functions 𝞥𝑔0 = 1𝑠 1 ∗ 1𝑠 2 = 𝑒 1 𝑒 2 𝑉𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦 ℰ= ≥ 𝐸𝑔
𝜋 𝜋 𝞥𝞥
1 2 1 2 𝑧 𝑧 1 1 2 1 2   −𝑧 −𝑧 1
𝐻 = − ∇1 − ∇2 − − + = − ∇1 − ∇2 − − + + +
2 2 𝑟1 𝑟2 𝑟12 2 2 𝑟1 𝑟2 𝑟1 𝑟2 𝑟12

3 −𝑟1 3 −𝑟2 1 2 1 2   𝑧− 𝑧− 1 3 −𝑟1 3 −𝑟2

ℰ= 𝑒 𝑒 − ∇ − ∇ − − − − +
𝜋 𝜋 2 1 2 2 𝑟1 𝑟2 𝑟1 𝑟2 𝑟12 𝜋 𝑒 𝜋 𝑒

3 −𝑟1 3 −𝑟2 1 2 1 2   3 −𝑟1 3 −𝑟2

= 𝑒 𝑒 − ∇ − ∇ − − -2
𝜋 𝜋 2 1 2 2 𝑟1 𝑟2 𝜋 𝑒 𝜋 𝑒 +

5
3 3 3 3
𝑒 −𝑟1 𝑒 −𝑟2 𝑒 −𝑟1 𝑒 −𝑟2
1
+ 8
𝜋 𝜋 𝑟12 𝜋 𝜋

3 −𝑟1 3 −𝑟2 𝑧 −  3 −𝑟1 3 −𝑟2 𝑁𝑒𝑒𝑑 𝑡𝑜 𝑏𝑒 𝑒𝑣𝑎𝑙𝑢𝑎𝑡𝑒𝑑

−2 𝜋 𝑒 𝜋 𝑒 𝑟1 𝜋 𝑒 𝜋 𝑒
 Helium Atom : Variation theory
Evaluation of electron nuclear attraction Optimizing ℰ
5
3 −𝑟1 3 −𝑟2 𝑧 −  3 −𝑟1 3 −𝑟2 ℰ = (−2 ) + − 2( 𝑧 −  )
𝜋 𝑒 𝜋 𝑒 𝑟1 𝜋 𝑒 𝜋 𝑒 8
5
= 2 − 2𝑧 +
8
3 −𝑟1 𝑧 −  3 −𝑟1 3 −𝑟2 3 −𝑟2
= 𝜋 𝑒 𝑟1 𝜋 𝑒 𝜋 𝑒 𝜋 𝑒 Setting
δℰ
= 0 𝑤𝑒 ℎ𝑎𝑣𝑒  = 𝑧 −
5
δ 16
1
27
2𝜋 𝜋 ∞
For z = 2 we have  =
16
3 −𝑟 𝑧 −  3 −𝑟 2
= න නන 𝑒 1 𝑒 1 𝑟1 𝑑𝑟1 sin 𝜃1 𝑑𝜃1 𝑑𝜑1 5
𝜋 𝑟1 𝜋 The Value of 16, is referred to as ‘Screening constant’
0 0 0
27
The Value of  = 16, referred as Effective nuclear charge.
3 ( − 𝑧) ∞ −2𝑟 𝜋 2𝜋
= න 𝑟1 𝑒 1 න sin 𝜃 𝑑𝜃 න
1 1 𝑑𝜑1
𝜋 0 0 0
Optimized ℰ is 77.48 eV (2% error)
1 4𝜋
2 2 From Table of integrals Further Improvements require additional variational
∞
𝑛! parameters Introduced either linear or non-linear form.
= 𝑧− න 𝑥 𝑛 𝑒 −𝑎𝑥 =
0 𝑎𝑛+1
Templates 𝑛 𝑛

෍ 𝐶𝑖 ψ𝑖 ෍ 𝐶𝑖 ψ𝑖
𝑖=1 𝑖=1
\bra
├෍ 𝑛 𝑛 𝑛 𝑛
𝑖=1 ෍ 𝐶 𝑖 ψ𝑖 ෍ 𝐶 𝑖 ψ𝑖 ඁ
𝑋 = \⟨෍ 𝐶𝑖 ψ𝑖 | ෍ 𝐶𝑖ψ𝑖 ۧ
𝑖=1 𝑖=1 𝑖=1 𝑖=1

𝑛 𝑛
𝑛 ෍ 𝐶 𝑖 ψ𝑖 𝐻 ෍ 𝐶 𝑖 ψ𝑖
ψ0𝑙 𝐻 ෍ 𝐶𝑖ψ𝑖 𝑖=1 𝑖=1
𝑖=1
\open is for grouping
\ for locking the size