In this chapter, we built vector addition and the scalar and vector products together with their interpretations as projections, areas and volumes, moved to index notation with the Kronecker delta and the Levi-Civita symbol to derive the vector triple product identity, and verified the results by explicit component calculations.
Elif Nur Gürbüz•November 15, 2026•18 min read
Mathematical Methods in Physics Chapter 1
Understanding a phenomenon in physics requires not only observing it but also knowing how to express it mathematically. Mathematical methods in physics are the collection of methods we use to define physical quantities, establish relationships between them, and solve the equations arising from these relationships. Ranging from vectors and differential equations to Fourier analysis and special functions, these methods help us see under which assumptions a problem can be solved and what the resulting solution means.
The development of these methods is closely connected to the questions posed by physics. Differential and integral calculus, developed by Newton and Leibniz in the seventeenth century, provided a strong foundation for studying motion and change. In the eighteenth and nineteenth centuries, work on mechanics, heat conduction, and electromagnetism fostered the development of the mathematical tools we use today.
In this series of articles, we will examine the fundamental mathematical methods used in physics step by step, beginning with vectors. Moving from definitions to derivations and examples, we will try to understand where the relations we use come from and which problems they help us solve.
Vectors
In this chapter, we will discuss vector addition, subtraction, and scalar and vector products. We will then examine the geometric meaning of triple products and see how vector identities are derived using the Kronecker delta and Levi-Civita symbols.
When stating the mass of an object, giving a single number together with an appropriate unit is sufficient. However, when describing the velocity of the same object, we also need information about direction. Moving east at five metres per second and moving north at the same speed are different motions.
We must preserve this distinction when expressing physical quantities mathematically. We represent quantities such as mass and temperature by scalars, and quantities such as displacement, velocity, and force by vectors.
Throughout this article, we will use a right-handed Cartesian coordinate system in three-dimensional Euclidean space. Our aim is to establish vector operations together with their geometric meanings.
Notation note. We will denote vectors with arrows above them, as in A, B, and their magnitudes without arrows, as in A=∣A∣. The unit vectors along the positive x,y,z axes will be ,,k, respectively. Ax, Ay, and Az are signed scalar components, not vectors.
How Do We Define a Vector?
Geometrically, we can represent a vector by a directed line segment. The length of the arrow indicates the magnitude of the vector, while the direction of the arrow indicates its orientation. Let us denote a vector by A and its magnitude by
A=∣A∣(1)
Here, A and A are not the same mathematical object. The former is a vector; the latter is a non-negative real number.
Vectors of magnitude one are called unit vectors. The unit vector in the direction of a non-zero vector A is given by
eA=∣A∣A(2)
Writing A=AeA in this way allows us to separate magnitude and direction. The zero vector has zero magnitude and no specific direction. Therefore, we cannot convert the zero vector into a unit vector.
Multiplication by a Scalar
For the product of a real number λ and a vector A, we have
∣λA∣=∣λ∣∣A∣(3)
If λ>0, the direction is preserved; if λ<0, it is reversed; if λ=0, the zero vector is obtained.
Example 1. Let us find the magnitude of the vector A=3−4 and the unit vector in its direction:
A=32+(−4)2=5,eA=AA=53−54.(4)
As a check, we find ∣eA∣=9/25+16/25=1. The unit vector preserves the direction of the original vector while reducing its magnitude to one.
Vector Addition and Subtraction
First, consider two vectors. If we start at a point and move by A, then by B, our total motion is
r=A+B(5)
Geometrically, we place the tail of B at the tip of A. The vector drawn from the initial starting point to the final point gives the sum. In this operation, the following properties hold:
A+B=B+A,(A+B)+C=A+(B+C)(6)
The types and units of the physical quantities being added must also be compatible. We can add two forces, but we cannot combine a force and a velocity in the same sum. We define subtraction through the opposite vector:
A−B=A+(−B).(7)
−B has the same magnitude as B and the opposite direction. Therefore,
A−B=−(B−A).(8)
Example 2. Let us walk three metres east and then four metres north. The total distance travelled is seven metres. The magnitude of our displacement is
∣r∣=32+42=5,m(9)
Here, the information about the direction of motion is what distinguishes distance travelled from displacement magnitude.
Multiplying Two Vectors
When extending multiplication from real numbers to vectors, we must specify what kind of object the result will be. In three dimensions, we use two fundamental operations: the scalar product and the vector product.
The Scalar Product of Two Vectors
Let θ be the angle between two vectors. We define
A⋅B=ABcosθ(10)
The result is a scalar. We can interpret this expression as a projection operation. The signed scalar component of A in the direction of B is Acosθ. Multiplying this by B gives the scalar product.
For two non-zero vectors, A⋅B=0 indicates that the angle between them is 90∘. Taking the scalar product of a vector with itself gives
A⋅A=A2(11)
Thus, we can calculate the magnitude of a vector using the scalar product:
∣A∣=A⋅A.(12)
Applied to the basis vectors, the scalar product of identical basis vectors gives 1, while that of different basis vectors gives 0:
⋅=0,⋅k=0,k⋅=0(13)⋅=⋅=k⋅k=1(14)
Angle and Projection Formulas
For A=0 and B=0, solving the definition of the scalar product for the angle gives
θ=arccos∣A∣∣B∣A⋅B,0≤θ≤π(15)
If the scalar product is positive, the angle is acute; if it is negative, the angle is obtuse.
For B=0, the signed scalar projection and vector projection of A in the direction of B are different objects:
The second formula is obtained by multiplying the first formula by the unit vector B/∣B∣. Thus, decomposing A into parts parallel and perpendicular to B, we can write
A∥=projectionBA,A⊥=A−A∥(17)
In this decomposition, A⊥⋅B=0 must hold.
Vector Product
The result of the vector product is a vector perpendicular to both of the vectors being multiplied. Its magnitude is given by
∣A×B∣=ABsinθ(18)
and its direction is determined by the right-hand rule. If the result is zero, it has no specific direction.
To see the geometric meaning of this magnitude, consider A and B as two sides of a parallelogram. Taking B as the base and Asinθ as the height, the area is
S=B(Asinθ)=∣A×B∣(19)
Reversing the order of the vector product reverses the direction:
A×B=−B×A.(20)
A direct consequence is the equality A×A=0. If the vector product of two non-zero vectors is zero, the vectors are parallel.
The vector product of each of the basis vectors ,,k with itself is the zero vector 0. The vector product of two different basis vectors gives the third basis vector, with a sign that depends on the order.
Rather than memorising these vector products of the basis vectors individually, we can make them easier to remember as follows: in the cycle →→k→, vector products taken in the cyclic order (×, ×k, k×) give a positive result, while products in the reverse order give a negative result. The result is always the third vector of the cycle.
Orthogonal Components: Making Geometry Calculable
Let us denote the unit vectors along the Cartesian axes by ,,k. We can write any vector as
A=Ax+Ay+Azk(25)
Here, Ax, Ay, and Az are signed scalar components. For example, Ay<0 tells us that the component of the vector along y points in the negative direction of that axis.
Since the unit vectors are mutually perpendicular,
∣A∣=Ax2+Ay2+Az2.(26)
If the angles made by the vector with the positive axes are α,β,γ, respectively, then
Ax=Acosα,Ay=Acosβ,Az=Acosγ.(27)
Adding the squares of these three equalities, for A=0, gives
cos2α+cos2β+cos2γ=1(28)
This means that we cannot choose the three direction cosines of a vector independently of one another.
Basic Operations in Cartesian Coordinates
When we resolve vectors into components, addition and subtraction are performed between components along the same axis:
A±B=(Ax±Bx)+(Ay±By)+(Az±Bz)k.(29)
For the scalar product, we use the perpendicularity of the unit vectors. The scalar product of unit vectors along different axes is zero, while that of unit vectors along the same axis is one.
and since their magnitudes are A=B=5, we find cosθ=2/5. Their vector product is
A×B=(2⋅1−0⋅0)+(0⋅2−1⋅1)+(1⋅0−2⋅2)k.(39)
that is, A×B=2−−4k. Let us check whether the result is indeed perpendicular to both vectors:
(2,−1,−4)⋅(1,2,0)(2,−1,−4)⋅(2,0,1)=0,=0.(40)
Therefore, a unit normal to the plane spanned by these two vectors is
n0=212−−4k(41)
The normal in the opposite direction is obtained by taking its negative.
Triple Products and Volume
One operation we can construct with three vectors is A⋅(B×C). This is called the scalar triple product.
First, consider B×C. The magnitude of this vector gives the base area, and its direction is perpendicular to the base. Taking its scalar product with A accounts for the signed height perpendicular to the base. Therefore, the geometric volume of the parallelepiped is
V=∣A⋅(B×C)∣(42)
The absolute value is important: the scalar triple product may be negative depending on orientation, but geometric volume cannot be negative.
Writing this in components,
A⋅(B×C)=AxBxCxAyByCyAzBzCz.(43)
Expanding this determinant along the first row gives
If the determinant is zero, the three vectors are linearly dependent; they cannot form a three-dimensional volume. Cyclic permutations do not change the result:
A⋅(B×C)=B⋅(C×A)=C⋅(A×B).(45)
Example 4. Let A=2, B=+3, and C=+4k:
B×CV=12−4+k,=∣2⋅12∣=24.(46)
If the components of the edge vectors are given in metres, the unit of volume is m3.
The other operation is the vector triple product, of the form A×(B×C). The result is a vector and can be simplified using the identity
A×(B×C)=B(A⋅C)−C(A⋅B)(47)
We must retain the parentheses here. The vector product is not associative; in general,
A×(B×C)=(A×B)×C.(48)
Let us now examine where this identity comes from.
Kronecker Delta and the Levi-Civita Symbol
From now on, let us denote the basis vectors by u1=, u2=, u3=k and the components by A1=Ax, A2=Ay, A3=Az:
A=A1+A2+A3k=i=1∑3Aiui.(49)
If we agree that an index repeated twice in a term is summed from 1 to 3 (the Einstein summation convention), we can drop the summation sign:
A=Aiui.(50)
The Kronecker delta is defined by
δij={1,i=j,\0,i=j,(51)
In an orthonormal basis, since ui⋅uj=δij, we can write
A⋅B=AiBjδij=AiBi(52)
The delta leaves only the contributions of matching indices in the sum. For example,
j=1∑3δ2jBj=0B1+1B2+0B3=B2.(53)
Under the summation convention, δii=δ11+δ22+δ33=3. We must remember that the repeated index i is summed over here.
Levi-Civita Symbol
In three dimensions, we define the Levi-Civita symbol by
ε123=ε231=ε312ε132=ε213=ε321=1,=−1(54)
Notice that if it is an even permutation of (1,2,3), the result is +1; if it is an odd permutation of (1,2,3), the result is −1.
If any two indices are equal, the symbol is zero. The ith component of the vector product is therefore
(A×B)i=εijkAjBk(55)
For example, let us choose i=1. Two non-zero terms remain:
Let us write our starting expression using indices:
[A×(B×C)]i=εijkAjεklmBlCm.(57)
Now let us use the identity
εijkεklm=δilδjm−δimδjl(58)
Substituting it, we obtain
[A×(B×C)]i=(δilδjm−δimδjl)AjBlCm.(59)
Expanding the parentheses, substituting i for l and j for m in the first term because i=l and j=m, and carrying out the calculation for the second term as well, we obtain:
=AjBiCj−AjBjCi(60)
Collecting matching indices together gives
=Bi(A⋅C)−Ci(A⋅B).(61)
Since this holds for every component, we obtain
A×(B×C)=B(A⋅C)−C(A⋅B).(62)
Index notation has allowed us to track which components match and which terms remain.
Checking the Same Result in Two Ways
Verifying a vector calculation using different methods is particularly useful for catching sign errors. Let us consider the following expression:
∣A×B∣2
From the geometric definition,
∣A×B∣2=A2B2sin2θ=A2B2(1−cos2θ).(63)
Since A⋅B=ABcosθ, we find
∣A×B∣2=A2B2−(A⋅B)2(64)
In our previous example,
∣A×B∣2=22+(−1)2+(−4)2=21.(65)
The right-hand side of the identity also gives (5)(5)−22=21. The component calculation and the geometric definition lead to the same result.
With the operations established in this chapter, we can now compare directions, take projections, find a normal to a plane, and calculate areas or volumes. In the next chapter of the series, we will move on to fields in which these vectors vary with position. The operations we have performed on constant vectors will thus form the foundation for differentiation and integration of fields.
Exercises
1. Find the magnitude of A=2−+2k, the unit vector in its direction, and its direction cosines. Show that cos2α+cos2β+cos2γ=1 holds.
[!note]- Answer A=4+1+4=3, eA=32−31+32k. The direction cosines are cosα=2/3, cosβ=−1/3, cosγ=2/3; the sum of their squares is 4/9+1/9+4/9=1.
2. Use the scalar product to find the angle between A=+ and B=+k.
[!note]- Answer A⋅B=1 and ∣A∣∣B∣=22=2, so cosθ=1/2 and θ=π/3=60∘.
3. Decompose A=2+3+k into components parallel and perpendicular to B=+. Verify that A⊥⋅B=0.
[!note]- Answer A⋅B=5 and ∣B∣2=2, so A∥=25(+) and A⊥=A−A∥=−21+21+k. Check: A⊥⋅B=−21+21+0=0.
4. Use the vector product to compute the area of the triangle with vertices P(1,0,0), Q(0,2,0), R(0,0,3).
[!note]- Answer PQ=−+2, PR=−+3k, and PQ×PR=6+3+2k. Its magnitude is 36+9+4=7, so the area of the triangle, half that of the parallelogram, is 7/2.
5. Do the vectors A=+2+3k, B=4+5+6k, C=7+8+9k form a parallelepiped? Decide using the scalar triple product.
[!note]- Answer A⋅(B×C)=1(45−48)−2(36−42)+3(32−35)=−3+12−9=0. The three vectors are linearly dependent (coplanar); the volume is zero, so they do not form a parallelepiped.
6. Using the Levi-Civita symbol and identity (58), show that (A×B)⋅(C×D)=(A⋅C)(B⋅D)−(A⋅D)(B⋅C). Setting C=A and D=B, recover relation (64).
[!note]- Answer (A×B)⋅(C×D)=εkijAiBj,εklmClDm. By cyclicity εkij=εijk, so (58) applies: (δilδjm−δimδjl)AiBjClDm=AiCiBjDj−AiDiBjCj, which is the required right-hand side. With C=A, D=B we get ∣A×B∣2=A2B2−(A⋅B)2.
7. For A=B= and C=, compute A×(B×C) and (A×B)×C to show that the vector product is not associative. Check the first result with identity (47).
[!note]- Answer A×(B×C)=×k=−, but (A×B)×C=0×=0. From the identity: B(A⋅C)−C(A⋅B)=(0)−(1)=−.