Group Theory in Particle Physics

Joshua Albert November 19, 2007

Group Theory

Group theory is a branch of mathematics which developed slowly over the years. Finding its origins in algebraic equations, number theory, and geometry, this eld would eventually be applied extensively to physics in the 20th century.

1.1

What is a Group?

A group is a set of objects with an associated operation which satises certain properties. For a group G which is a set of elements {T, T , T , ...} with operator , we require: 1. Closure. T, T G : T T G . This property of closure should be familiar to anyone with a linear algebra background, as a vector space also fullls this property, with addition as the operator. 2. Associativity. T, T , T G : T (T T ) = (T T ) T . This is just the associative property which we often nd in regular algebra. Note that this does not imply the commutative property. 3. Identity. I G such that T G : T I = I T = T . Thus, there exists an element I, called the identity element in any group. This property is useful in dening the next: 4. Inverse. For each T GT 1 such that T T 1 = T 1 T = I . Any set with an operator satisfying these four properties is a group. For further examples, the operator may not be explicitly shown, but rather the implied multiplication operation will represent whatever operation with respect to which the group is dened. The group operation is often referred to as group multiplication.

Of course, the way we dened a group is very abstract. We can use dierent operators and have dierent types of objects as our group elements. Here are a few examples: 1. The set of integers under addition. The set of integers Z satises all the requirements of a group under addition. Addition of two integers will never produce a non-integer, addition is associative, the integer 0 is the identity element, and every element has an inverse, specically that element multiplied (in the traditional sense) by -1. 2. The set of real numbers under multiplication. The set of real numbers R is a group under multiplication. It is closed and associative, 1 is the identity and 1/x is the inverse of x. It should also be noted that R is also a group under addition. 3. Parity. One can dene a group consisting of two elements, one positive, one negative (or even and odd). They can be represented at +1 and -1 and combine under the multiplication operation. The parity group is an example of a simple group which ends up being used often in physics. Something else worth noting is that while the previous two groups had innite numbers of elements, this group has a nite number of elements. The number of elements in a group is its order. 4. Rotations in space. In the group of rotations in space, each element is a dierent rotation, which can be specied by the three Euler angles. The operation for this group is the composition operation, that is, performing one rotation, then the next. Obviously, two rotations put together are just another rotation of the system. Associativity is maintained, the identity is the null rotation, and any rotation clearly has an inverse to cancel it out. It should be noted that this group, unlike the other ones mentioned so far, does not have commutivity as a property. That is, for two rotations R and R , RR = R R. A group with non-commuting elements is called non-Abelian. Not surprisingly, a group with entirely commuting elements, like the previous three examples, is called an Abelian group. Thus, we have seen examples of groups with very dierent kinds of elements, and other groups can be imagined. For example, a group of permutations on a Rubrics cube, rearrangements of letters in a word, or the set of unitary matrices. In its most abstract form, a nite group can be represented simply as a multiplication table which indicates all the possible multiplications. Consider the group of the powers of the imaginary constant i. A multiplication table for this group is shown in Table 1. While here the group elements are labelled intuitively with values, an equivalent group can be represented using arbitrary labels, such as A, B, C, D instead of i, 1, i, 1. These tables are the most abstract way of representing a group, but certainly not the most useful. 2

Table 1: Multiplication Table for Powers of i i 1 i 1 i 1 i 1 i 1 i 1 1 i i 1 i 1 i 1 i 1 i 1

1.2

Representations

Clearly, groups range in type from very abstract, to very concrete. A good way of representing group elements is needed to make things more managable. That is where matrix representations come in. Each element in a group can be assigned a n n matrix to represent it, provided it meets one condition. If each element T G is represented by (T ), we must have (T T ) = (T )(T ) for each T, T G . If this condition is met, then is a d-dimensional representation of G . Note that matrix multiplication is always used as the group multiplication operation. Now that we can look at groups in a more concrete way, this is a good time to discuss some group properties. First of all, using matrices to represent group elements has some clear advantages, since matrices, like non-Abelian group elements, do not necessarily commute. Matrices also automatically have the property of being associative, and the identity matrix is already dened. The existence of an inverse group element simply requires that all our group elements be represented by square non-singular matrices. Equivalent representations are representations which are related by a similarity tranfsormation. For a xed operator (a constant matrix), and T G , (T ) = (T )1 and (T ) and (T ) are equivalent representations. It is easily demonstrated that the new representation is still valid:

(T1 ) (T2 ) = (T1 )1 (T2 )1 = (T1 )(T2 )1 = (T1 T2 )1 = (T1 T2 )

By using similarity transformations, we can change one representation into a potentially more useful one. An important concept in representations is the idea of reducibility. If, by similarity transformation, a group representation can be converted to block diagonal form, it is a reducible representation. If such a transformation is not possible, it is an irreducible representation. This brings up one more group concept the direct product of groups. The direct product is not exclusive the realm of representations; on the contrary, it is a property of the groups themselves. Essentially, two groups are combined in a direct product () to produce elements which have a component from each group. These elements operate on each other as follows: For Ti group G , and Sj group H: Uij = Ti Sj G H Uij Ukl = (Ti Tj ) (Sk Sl ) The representation for a direct product group is a block diagonal matrix, with each block as one of the irreducible representations. Since dierent blocks in block diagonal matrices do not interact through multiplication, and multiply normally with themselves, this is a good representation.

1.3

Some Examples of Useful Groups

Most of the useful nite groups have already been described in section 1.1. In fact, the main useful nite group to be described in this report is the parity group. This is related to the parity conservation which is so useful in calculating possible transitions and so on in nuclear physics. Parity is the Z2 group. The groups which are most important for our purposes are Lie groups. A Lie group is dened as a group in which the elements are labelled by a set of continuous parameters with a multiplication law that depends smoothly on the parameters. [2] These Lie groups can describe many continuous symmetries, and, once Lie algebras are developed, can describe some discrete phenomena too. The rst group of note is the set of all rotations in 3-dimensional space. This group is called O(3). It is represented by the group of all 3-dimensional real orthogonal matrices. Because the matrix entries should be real, this group is sometimes referred to by O(3, R). This group describes both proper and improper rotations, so not every element of the group is connected smoothly to every other element. Any improper rotation can be represented as a product of a proper rotation and an inversion of spatial coordinates. Of course, if all proper rotations form a group of their own, then this group can also be identied. This group is called SO(3), or SO(3, R). This is the set of all 3 3 orthogonal matrices with determinant 1 (the 4

special orthogonal group). Both the orthogonal and special orthogonal groups can be generalized to a higher (or lower) number of dimensions. For example, SO(4, R) is the set of all 4 4 orthogonal matrices with determinant 1. In general, groups SO(n, R) and O(n, R) can be constructed for any n 0, although for the lowest values of n these are trivial. Another similar set of groups are the unitary (U (n)) and special unitary (SU (n)) groups. These consist of all unitary n n matrices and all unitary n n matrices with determinant 1, respectively. There is no restriction that the entries be real. It is not immediately obvious that these have a physical signcance. However, the (special) unitary groups can have close ties to the (special) orthogonal groups. For example, consider U (1), which is the group of complex numbers with unit modulus. This group is actually isomorphic to SO(2, R) [1], and each describe rotations in a plane. To say that two groups are isomorphic means that there is a one-to-one mapping from one group to the other satisfying

(T1 )(T2 ) = (T1 T2 )

for (T ) the isomorphic mapping of group elements Ti . Similarly, a homomorphic (like isomorphic, but not one-to-one) mapping exists between SU (2) and SO(3). Other useful groups are best introduced by their physical manifestations. The Lorentz group is the group of all 3-space rotations and Lorentz boosts. The Lorentz group denes a set of transformations under which physical laws must be invariant (lest they violate special relativity) but which leaves one point invariant (specically, the origin). This is a 6-dimensional Lie group, and the denition of the dimensionality of a Lie group will be described in section 2.2. When combined with the 4-dimensional Lie group of translations in space and time (x,y,z,t), the result is the Poincar e group, the full symmetry group of relativistic eld theory [5].

Lie Groups and Symmetries

Lie groups are a very special class of continuous groups. In this report only classical Lie groups will be considered; this restriction allows us to examine some fascinating properties of Lie groups. Through these groups we will nd connections to particle physics.

2.1

Lie Algebras

As described before, a Lie Group has elements which are dened by continuous parameters. It is possible to construct an element of a Lie group by considering how far away from the identity element the given element is. This is somewhat confusing, but it can be made fairly clear with an example. Consider the Lie group of translations in one dimension [5](this is easily generalized to more dimensions). The translation operator operates on a position dened state so that:

T (x)|x0 = |x + x0

(1)

We can dene a generator of translation P (the symbol is not random) such that an innitesimal displacement is given by: T (dx) I idxP Our group properties let us write: T (x + dx) = T (x) + dx and T (x + dx) = T (dx)T (x) Then, substituting 2 into 4 and equating to 3 we get dT (x) = iP T (x) dx which solves to: T (x) = eiP x (6) (4) dT (x) dx (3) (2)

(5)

We see that the translation operator can be written as an exponentiation of innitesimal translations. In quantum physics, the generator of translations, P, is actually the momentum operator. In general, any simple Lie group can be dened as an exponentiation of innitesimal generators, the generators being known as Lie algebras. Lie algebras have their own dening properties, and much can be learned about a Lie group by studying the algebras which generate it. Lie algebras and their representations have equivalence relations just like the groups. The dimension of a Lie group is given by the number of independent Lie algebras which generate it.

Thus, the group of all translations in 3-dimensional space is 3-dimensional, as it is generated by the three translation operators for the three cartesian directions.

2.2

Continuous Symmetries

The translation operator brings up a symmetry immediately: translation symmetry. All physical laws are equivalent if the coordinate axes are shifted by some spatial translation. This is represented by the Lie group of translation operators. It is trivial to check that this group obeys all four rules for being a group. Moreover, because this is a continuous symmetry, Noethers theorem requires that there be a conservation law associated with it. In this case, it is conservation of linear momentum. This pattern of the generators being represented by the operators for a conserved quantity can be found in rotational symmetry as well. The angular momentum operators Lx , Ly and Lz are the Lie algebras for the rotation group. The spin one angular momentum operators (for SO(3)) have the same commutation structure as the SU (2) generators, which are the Pauli matrices. While the Pauli matrices always exponentiate to SU (2), constructing SO(3) can be a little bit tricky [1], and will not be treated here. Finally, the time translation operator is the Hamiltonian, and this leads to the connection between time invariance and conservation of energy. Now the Poincar e groups denition as being a 10-dimensional Lie group is clear. Three algebras correspond to the three generators of rotations, and three more correspond to the generators of boosts. These six are the algebras of the Lorentz group. Then, the three translation operators and the time translation operator ll out the remaining four Lie algebras.

Particles and Forces

Lie groups and Lie algebras have enough depth of complexity to study in great detail, far beyond the scope of this report. The concepts of classes, weights, roots, and so on can be useful in this analysis, but are also, in general, beyond the scope of this report. We will, however, examine SU (3) in detail, as this group plays a very important role in particle physics. We will nd one inexact symmetry in SU (3) (the up-down-strange quarks yielding the Eightfold Way), and one exact symmetry (the QCD representation of gluons).

3.1

Taming the Particle Zoo (Approximate SU (3))

The generators of SU (3) can be written in dierent bases, but the standard one is the Gell-Mann matrices. 0 1 1 = 1 0 0 0 1 0 3 = 0 1 0 0 0 0 5 = 0 0 i 0 0 0 7 = 0 0 0 i 0 0 2 = 0 i 0 0 0 0 4 = 0 0 1 0 i 0 6 = 0 0 0 0 0 1 1 i 8 = 3 0 0 0 i 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 0 2

The generators we want are actually given by Ti = i /2. It should be noted that there are only eight generators, despite the appearance that there would be nine degrees of freedom (from nine matrix entries). However, a property of Lie algebras is that they must be traceless, and this eectively eliminates one degree of freedom, thus yielding only eight generators. We note that T1 , T2 and T3 form a subalgebra of SU (2). This generates the isospin subgroup. Taking T3 and T8 , the two diagonal elements, we get what is called a Cartan subalgebra [2]. As diagonal matrices, they will naturally have the three cartesian unit vectors as eigenvectors. In the |H1 , H2 basis, for H1 = T3 and H2 = T8 , we nd: 1 1 0 corresponds to | 1 , 2 2 3 0 0 1 1 corresponds to | 1 , 2 2 3 0 8

 0 1 0 corresponds to |0, 3 1 Plotting H2 vs. H1 gives three points in an equilateral triangle. These points correspond to the up, down, and strange quark. The H1 number corresponds to the z-component of isospin, I3 , and H2 = 3Y /2 for Y the hypercharge, dened by the sum of the baryon number and strageness (strangely, a strange quark has strangeness -1). Now, from here, we can easily begin producing combinations of these quarks, and plotting them on Y vs. I3 . For the mesons, see gure 1.

s=1  T 

0 K ) r (ds 

s=0

 (du ) r T

0 r r T Tr K (su ) q = 1

s = 1

+ ) rK (us T T T ) T r + (ud   q=1   r 0 ) K (sd

q=0 [6]). 0 is given by

Figure 1:

A The meson octet (used L TEX code from 2 ss is given by ( 1 + dd )). 6 (uu

)) and dd

Other, similar pictures can be produced for baryons as well. Because baryons have three quarks, the group involved is actually 3 3 3 = 10 8 8 1 giving us the 27 elements we should expect from all the permutations of the three quark avors in three type positions [4]. However, this representation of SU(3) is reduced into four irreducible representations, each of which is a group by itself. Just like the mesons t into the octet, there is a baryonic decouplet, two octets, and a singlet. A nice feature of these organizations, commonly referred to as the eightfold-way, is that the particles can be placed into these arrangements based on their physical properties alone. While organization by hypercharge and isospin is exact, there is also an approximate pattern in the masses. Because the up and down quarks are so similar in mass, and because the strange quark is also a light quark, the rows of particles in these arrangements tend to have similar masses. The baryon decouplet is shown in gure 2. The SU (3) symmetry exhibited in generating the Eightfold Way is a result of the approximate symmetry

s=0

s = 1

s = 2 s = 3

(ddd) 0 (ddu) + (duu) ++ (uuu) r r r r T  T  q=2 T  T r (dds) r 0 (uds) r + (uus) T  T  q=1 T  T r  r 0 T (dss)  (uss) T  T  q=0 T  Tr  (sss)

q = 1

A Figure 2: The baryon decouplet (used L TEX code from [6]). Especially notable is the , which was predicted using the Eightfold Way by Gell-Mann before it was discovered [3].

between the three light quarks, up, down and strange. Isospin is closer to an exact symmetry because the up and down quarks have nearly the same mass. This model is not very useful when expanded to SU(4) with the charm quark, because its mass is much greater than that of the strange quark, and much much greater than that of the up or down quark. Thus, trying to unite all six quarks under a SU(6) symmetry can be misleading, as this is not an exact symmetry. However, the Eightfold Way can provide a convenient and somewhat intuitive organization method for hadronic particles. However, there is nothing unique about this representation, it is merely a consequence of the quarks which make up the baryons and mesons.

3.2

Constructing The Force Carriers: SU (3) and Beyond

While SU (3) is an approximate avor symmetry due to quark mass dierences, it is an exact color symmetry! The symmetry SU (3) is in fact the basis for QCD, and its 8-dimensional generator basis (the Gell-Mann matrices) is reected in the octet of gluons. While all combinations of red, green and blue with their anticolors should give nine states, one of them is a color singlet, and would have some non-gluon like properties such as a long range, and thus it is believed that this does not exist (at least as a gluon) [3]. The color octet

10

is: |1 = (r b + br )/ 2 |3 = (rr b b)/ 2 |5 = i(rg gr )/ 2 |7 = i(bg g b)/ 2 and the color singlet is: |9 = (rr + b b + gg )/ 3 Gluons are the carriers of the strong force, and the strong force is described by SU(3) symmetry. Of course, attempts have been made to associate the other forces with groups. There has actually been considerable success here, as the weak and electromagnetic forces are actually united in this way. The electroweak interaction is based on a combined Lie algebra of U (1) SU (2). Salam, Weinberg and Glashow were awarded the 1979 Nobel Prize in Physics for their contributions to unifying electromagnetism and the weak interaction. Attempts to unify the strong force, and even gravity with the electroweak into one single interaction are underway. It is major goal of theoretical physics to show that all the forces are in fact manifestations of a single phenomenon. Group theory, with its organization of like elements into recognizable patterns, may be an important part of achieving this goal. |2 = i(r b br )/ 2 |4 = (rg + gr )/ 2 |6 = (bg + g b)/ 2

|8 = (rr + b b 2g g )/ 6

References
[1] J.F. Cornwell. Group Theory in Physics: An Introduction. Academic Press, Cambridge, UK, 1997. [2] Howard Georgi. Lie Algebras in Particle Physics. The Benjamin/Cummings Publishing Company, Massachusetts, USA, 1982. [3] David Griths. Introduction to Elementary Particles. John Wiley and Sons, Inc., New York, USA, 1987. [4] Francis Halzen and Alan D. Martin. Quarks and Leptons: An Introductory Course in Modern Particle Physics. John Wiley and Sons, New York, USA, 1984. [5] Wu-Ki Tung. Group Theory in Physics. World Scientic, Philadelphia, USA, 1985.

11

[6] Eightfold way (physics). http://en.wikipedia.org/wiki/Eightfold_way_(physics)

12