Masteries · Jason GlowneyArithmoí Book 1

Masteries

1.1.20 The U(3)U(3) Symmetry Group

We are now able to build on the S3ℂS_{3\mathbb{C}} sym­me­try group and what we have learned about geo­met­ric sym­me­try rep­re­sen­ta­tion in ℂ3.\mathbb{C}^{3}.The bound­aries of the S3ℂS_{3\mathbb{C}} group are rep­re­sent­ed by three in­ter­con­nect­ed com­plex cir­cles, which act to cir­cum­scribe the equi­lat­er­al tri­an­gle's ver­tices. If we define the dis­tance from the origin as lj,l_{j},each vertex of the equi­lat­er­al tri­an­gle will be a unit mod­u­lus from the origin

|lj|=lj∙lj∗=1.\begin{array}{r} \left| l_{j} \right| = \sqrt{l_{j} \bullet l_{j}^{*}} = 1. \end{array}
(1.1.96)
Figure
Figure 1.1.10 – The equi­lat­er­al tri­an­gle bound­ed by com­plex unit cir­cles ∈ℂ3\in \mathbb{C}^{3}.

We can also see that each of the ver­tices are sep­a­rat­ed by edge lengths equal to |2|\left| \sqrt{2} \right|, as de­pict­ed in Figure 1.1.10. With the geom­e­try of the com­plex planes set to unit lengths at the bound­aries along the axes, and the two-di­men­sion­al area of the equi­lat­er­al tri­an­gle uni­tar­i­ly set as 12(2∙2)=1,\frac{1}{2}\left( \sqrt{2} \bullet \sqrt{2} \right) = 1,we can better con­sid­er the geom­e­try ∈ℂ3.\in \mathbb{C}^{3}.In this way, each com­plex unit circle plane can be rep­re­sent­ed in polar form as

|eiX1|, |eiX2|,|eiX3|=1,\begin{array}{r} \left| e^{iX_{1}} \right|,\ \left| e^{iX_{2}} \right|,\left| e^{iX_{3}} \right| = 1, \end{array}
(1.1.97)

where the mod­u­lus rr is omit­ted, since it is equal to 1. For geo­met­ric clar­i­ty, we name the imag­i­nary-valued per­mu­ta­tions rep­re­sen­ta­tive of ℂ3\mathbb{C}^{3}, using XjX_{j}, which is rep­re­sen­ta­tive of tri­an­gle's ver­tices in each com­plex plane, seen as

X1=(i0000i0i0),  X2=(00i0i0i00),  X3=(0i0i0000i).\begin{array}{r} X_{1} = \begin{pmatrix} i & 0 & 0 \\ 0 & 0 & i \\ 0 & i & 0 \end{pmatrix},\ \ X_{2} = \begin{pmatrix} 0 & 0 & i \\ 0 & i & 0 \\ i & 0 & 0 \end{pmatrix},\ \ X_{3} = \begin{pmatrix} 0 & i & 0 \\ i & 0 & 0 \\ 0 & 0 & i \end{pmatrix}. \end{array}
(1.1.98)

The real-valued el­e­ments of the group are de­not­ed by YjY_{j}, with Y1Y_{1} being the iden­ti­ty el­e­ment ϵ\epsilon

Y1=(100010001),  Y2=(010001100),  Y3=(001100010).\begin{array}{r} Y_{1} = \begin{pmatrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1 \end{pmatrix},\ \ Y_{2} = \begin{pmatrix} 0 & 1 & 0 \\ 0 & 0 & 1 \\ 1 & 0 & 0 \end{pmatrix},\ \ Y_{3} = \begin{pmatrix} 0 & 0 & 1 \\ 1 & 0 & 0 \\ 0 & 1 & 0 \end{pmatrix}. \end{array}
(1.1.99)

We define an al­ge­bra 𝔤\mathfrak{g}, as the set of el­e­ments

𝔤={Xj,Yj}\begin{array}{r} \mathfrak{g} =\left\{ X_{j},Y_{j} \right\} \end{array}
(1.1.100)

The­o­rem (1.1.101) – The set of per­mu­ta­tions 𝔤∈U(3)\mathfrak{g} \in U(3) forms a group under com­po­si­tion.

Proof –

  • Clo­sure → ∀ X,Y∈𝔤,  X ∙Y∈𝔤,\rightarrow \ \forall\ X,Y \in \mathfrak{g},\ \ X\ \bullet Y \in \mathfrak{g},
  • As­so­cia­tiv­i­ty → ∀ X,Y,Z∈G,  ( X ∙Y)∙Z=X∙(Y∙Z),\rightarrow \ \forall\ X,Y,Z \in G,\ \ (\ X\ \bullet Y) \bullet Z = X \bullet (Y \bullet Z),
  • Iden­ti­ty → ∃ e∈𝔤:∀ a∈G,  a∙e=e∙a=a,\rightarrow \ \exists\ e \in \mathfrak{g}:\forall\ a \in G,\ \ a \bullet e = e \bullet a = a,
  • In­verse El­e­ment →∀ a∈G ∃ b∈G: a∙b=b∙a=e.\rightarrow \forall\ a \in G\ \exists\ b \in G:\ a \bullet b = b \bullet a = e.
 □\begin{array}{r} \ \square \end{array}

De­f­i­n­i­tion (1.1.102) – A finite-di­men­sion­al real or com­plex Lie al­ge­bra is a finite-di­men­sion­al real or com­plex vector space 𝔤\mathfrak{g}, to­geth­er with a Lie brack­et [X,Y]\lbrack X,Y\rbrack, from 𝔤×𝔤\mathfrak{g} \times \mathfrak{g} onto 𝔤\mathfrak{g}, having the fol­low­ing prop­er­ties1

  • Bi­lin­ear­i­ty:
[aX+bY,Z]=a[X,Z]+b[Y,Z] and [Z,aX+bY]=a[Z,X]+b[Z,Y] ∀ X,Y,Z ∈ 𝔤,\lbrack aX + bY,Z\rbrack = a\lbrack X,Z\rbrack + b\lbrack Y,Z\rbrack\ and\ \lbrack Z,aX + bY\rbrack = a\lbrack Z,X\rbrack + b\lbrack Z,Y\rbrack\ \forall\ X,Y,Z\ \in \ \mathfrak{g},

and for ar­bi­trary num­bers aa and bb.

  • Skew-sym­met­ric:
[X,Y]=−[Y,X],  [X,Z]=−[Z,X],  [Y,X]=−[X,Y], \begin{array}{r} \lbrack X,Y\rbrack = - \lbrack Y,X\rbrack,\ \ \lbrack X,Z\rbrack = - \lbrack Z,X\rbrack,\ \ \lbrack Y,X\rbrack = - \lbrack X,Y\rbrack,\ \end{array}
  • The Jacobi iden­ti­ty holds:
[X,[Y,Z]]+[Z,[X,Y]]+[Y,[Z,X]]=0, ∀ X,Y,Z ∈ 𝔤.\begin{array}{r} \left\lbrack X,\lbrack Y,Z\rbrack \right\rbrack + \left\lbrack Z,\lbrack X,Y\rbrack \right\rbrack + \left\lbrack Y,\lbrack Z,X\rbrack \right\rbrack = 0,\ \forall\ X,Y,Z\ \in \ \mathfrak{g}. \end{array}

The­o­rem (1.1.103) – The set of per­mu­ta­tions 𝔤={Xj,Yj}\ \mathfrak{g} =\left\{ X_{j},Y_{j} \right\} forms a finite-di­men­sion­al com­plex Lie al­ge­bra over a com­plex vector space.

Proof – The Lie Al­ge­bra De­f­i­n­i­tion

  • Bi­lin­ear­i­ty: [aX+bY,Z]=a[X,Z]+b[Y,Z]\ \lbrack aX + bY,Z\rbrack = a\lbrack X,Z\rbrack + b\lbrack Y,Z\rbrackand [Z,aX+bY]=a[Z,X]+b[Z,Y] ∀ X,Y,Z ∈\lbrack Z,aX + bY\rbrack = a\lbrack Z,X\rbrack + b\lbrack Z,Y\rbrack\ \forall\ X,Y,Z\ \in 𝔤\mathfrak{g}

and for ar­bi­trary num­bers a and ba\ and\ b,

  • Skew-sym­met­ric:

[X,Y]=−[Y,X] ∀\lbrack X,Y\rbrack = - \lbrack Y,X\rbrack\ \forall X,Y∈𝔤X,Y \in \mathfrak{g},

  • The Jacobi Iden­ti­ty holds:

[X,[Y,Z]]+[Y,[Z,X]]+[Z,[X,Y]]=0,\left\lbrack X,\lbrack Y,Z\rbrack \right\rbrack + \left\lbrack Y,\lbrack Z,X\rbrack \right\rbrack + \left\lbrack Z,\lbrack X,Y\rbrack \right\rbrack = 0, ∀ X,Y∈𝔤\forall\ X,Y \in \mathfrak{g}.

□\square

Taking the ex­po­nen­tial map­ping of our Lie al­ge­bra el­e­ments ∈𝔤\in \mathfrak{g}

G=e𝔤={eXj,eYj},\begin{array}{r} G = e^{\mathfrak{g}} = \left\{ e^{X_{j}},e^{Y_{j}} \right\}, \end{array}
(1.1.104)

we can define that a matrix Lie group by ex­po­nen­ti­a­tion is a smooth man­i­fold over GG.

De­f­i­n­i­tion (1.1.105) – The Uni­tary Group 𝐔(𝟑)\mathbf{U}\left( \mathbf{3} \right) is a group under com­po­si­tion, sit­ting in the com­plex geo­met­ric space ℂ3×3\mathbb{C}^{3 \times 3}

U(3)={G∈ℂ3×3:G∗G=I,|det⁡(G)|=I},\begin{array}{r} U(3) = \left\{ G \in \mathbb{C}^{3 \times 3}:G^{*}G = I,\left| \det(G) \right| = I \right\}, \end{array}

with a set of Lie al­ge­bra el­e­ments 𝔤={Xj,Yj}\mathfrak{g} =\left\{ X_{j},Y_{j} \right\}, for angle ar­gu­ments xj,yj∈ℝx_{j},y_{j}\in \mathbb{R}

𝔤={(ix10000ix10ix10),   (00ix20ix20ix200),   (0ix30ix30000ix3),(y1000y1000y1),   (0y2000y2y200),   (00y3y3000y30)}.\begin{array}{r} \mathfrak{g} =\left\{ \begin{array}{r} \begin{pmatrix} ix_{1} & 0 & 0 \\ 0 & 0 & ix_{1} \\ 0 & ix_{1} & 0 \end{pmatrix},\ \ \ \begin{pmatrix} 0 & 0 & ix_{2} \\ 0 & ix_{2} & 0 \\ ix_{2} & 0 & 0 \end{pmatrix},\ \ \ \begin{pmatrix} 0 & ix_{3} & 0 \\ ix_{3} & 0 & 0 \\ 0 & 0 & ix_{3} \end{pmatrix}, \\ \begin{pmatrix} y_{1} & 0 & 0 \\ 0 & y_{1} & 0 \\ 0 & 0 & y_{1} \end{pmatrix},\ \ \ \begin{pmatrix} 0 & y_{2} & 0 \\ 0 & 0 & y_{2} \\ y_{2} & 0 & 0 \end{pmatrix},\ \ \ \begin{pmatrix} 0 & 0 & y_{3} \\ y_{3} & 0 & 0 \\ 0 & y_{3} & 0 \end{pmatrix} \end{array} \right\}. \end{array}
(1.1.106)

The set of ro­ta­tion ma­tri­ces G={eXj,eYj}G = \left\{ e^{X_{j}},e^{Y_{j}} \right\}, in polar form, can take both ± ∠ arguments∈ℝ\pm \ \angle\ arguments\in \mathbb{R}, seen for eXje^{X_{j}}, as

GX1:{GX1(x1)=(eix1000cos⁡(x1)isin(x1)0isin(x1)cos⁡(x1)),  GX1(−x1)=(e−ix1000cos⁡(x1)−isin(x1)0−isin(x1)cos⁡(x1))},\begin{array}{r} {G_{X}}_{1}:\left\{ {G_{X}}_{1}\left( x_{1} \right) = \begin{pmatrix} e^{ix_{1}} & 0 & 0 \\ 0 & \cos\left( x_{1} \right) & isin\left( x_{1} \right) \\ 0 & isin\left( x_{1} \right) & \cos\left( x_{1} \right) \end{pmatrix},\ \ {G_{X}}_{1}\left( {- x}_{1} \right) = \begin{pmatrix} e^{- ix_{1}} & 0 & 0 \\ 0 & \cos\left( x_{1} \right) & - isin\left( x_{1} \right) \\ 0 & - isin\left( x_{1} \right) & \cos\left( x_{1} \right) \end{pmatrix} \right\}, \end{array}
(1.1.107)
GX2:{GX2(x2)=(cos⁡(x2)0isin(x2)0eix20isin(x2)0cos⁡(x2)),  GX2(−x2)=(cos⁡(x2)0−isin(x2)0e−ix20−isin(x2)0cos⁡(x2))},\begin{array}{r} {G_{X}}_{2}:\left\{ {G_{X}}_{2}\left( x_{2} \right) = \begin{pmatrix} \cos\left( x_{2} \right) & 0 & isin\left( x_{2} \right) \\ 0 & e^{ix_{2}} & 0 \\ isin\left( x_{2} \right) & 0 & \cos\left( x_{2} \right) \end{pmatrix},\ \ {G_{X}}_{2}\left( {- x}_{2} \right) = \begin{pmatrix} \cos\left( x_{2} \right) & 0 & - isin\left( x_{2} \right) \\ 0 & e^{- ix_{2}} & 0 \\ - isin\left( x_{2} \right) & 0 & \cos\left( x_{2} \right) \end{pmatrix} \right\}, \end{array}
(1.1.108)
GX3:{GX3(x3)=(cos⁡(x3)isin(x3)0isin(x3)cos⁡(x3)000eix3),  GX3(−x3)=(−cos⁡(x3)−isin(x3)0isin(x3)cos⁡(x3)000e−ix3)}.\begin{array}{r} {G_{X}}_{3}:\left\{ {G_{X}}_{3}\left( x_{3} \right) = \begin{pmatrix} \cos\left( x_{3} \right) & isin\left( x_{3} \right) & 0 \\ isin\left( x_{3} \right) & \cos\left( x_{3} \right) & 0 \\ 0 & 0 & e^{ix_{3}} \end{pmatrix},\ \ {G_{X}}_{3}\left( {- x}_{3} \right) = \left( - \begin{matrix} \cos\left( x_{3} \right) & - isin\left( x_{3} \right) & 0 \\ isin\left( x_{3} \right) & \cos\left( x_{3} \right) & 0 \\ 0 & 0 & e^{- ix_{3}} \end{matrix} \right) \right\}. \end{array}
(1.1.109)

The non-iden­ti­ty cyclic per­mu­ta­tion el­e­ments take on a more com­pli­cat­ed form with ex­po­nen­ti­a­tion. Ef­fec­tive­ly, the ex­po­nen­tial series eixe^{ix} is split three ways in­stead of the cus­tom­ary two-way splits we know in the trigono­met­ric form by the sine and cosine func­tions. We will study these three-way trigono­met­ric func­tions in more depth in later sec­tions. To be com­plete, we present the Lie groups GY2{G_{Y}}_{2},GY3{G_{Y}}_{3}, as fol­lows:

GY2:{GY2(y2)=(ey23+23e−y22cos⁡(3y22)ey23−23e−y22sin⁡(π6+3y22)ey23−23e−y22sin⁡(π6−3y22)ey23−23e−y22sin⁡(π6−3y22)ey23+23e−y22cos⁡(3y22)ey23−23e−y22sin⁡(π6+3y22)ey23−23e−y22sin⁡(π6+3y22)ey23−23e−y22sin⁡(π6−3y22)ey23+23e−y22cos⁡(3y22)),GY2(−y2)=(e−y23+23ey22cos⁡(3y22)e−y23−23ey22sin⁡(π6−3y22)e−y23−23ey22sin⁡(π6+3y22)e−y23−23ey22sin⁡(π6+3y22)e−y23+23ey22cos⁡(3y22)e−y23−23ey22sin⁡(π6−3y22)e−y23−23ey22sin⁡(π6−3y22)e−y23−23ey22sin⁡(π6+3y22)e−y23+23ey22cos⁡(3y22))}(1.1.110)\begin{array}{r} G_{Y_{2}}:\left\{ \begin{array}{r} G_{Y_{2}}\left( y_{2} \right) = \begin{pmatrix} \frac{e^{y_{2}}}{3} + \frac{2}{3}e^{- \frac{y_{2}}{2}}\cos\left( \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{y_{2}}}{3} - \frac{2}{3}e^{- \frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{y_{2}}}{3} - \frac{2}{3}e^{- \frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{2}}{2} \right) \\ \frac{e^{y_{2}}}{3} - \frac{2}{3}e^{- \frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{y_{2}}}{3} + \frac{2}{3}e^{- \frac{y_{2}}{2}}\cos\left( \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{y_{2}}}{3} - \frac{2}{3}e^{- \frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{2}}{2} \right) \\ \frac{e^{y_{2}}}{3} - \frac{2}{3}e^{- \frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{y_{2}}}{3} - \frac{2}{3}e^{- \frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{y_{2}}}{3} + \frac{2}{3}e^{- \frac{y_{2}}{2}}\cos\left( \frac{\sqrt{3}y_{2}}{2} \right) \end{pmatrix}, \\ G_{Y_{2}}\left( {- y}_{2} \right) = \begin{pmatrix} \frac{e^{{- y}_{2}}}{3} + \frac{2}{3}e^{\frac{y_{2}}{2}}\cos\left( \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{- y_{2}}}{3} - \frac{2}{3}e^{\frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{{- y}_{2}}}{3} - \frac{2}{3}e^{\frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{2}}{2} \right) \\ \frac{e^{{- y}_{2}}}{3} - \frac{2}{3}e^{\frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{{- y}_{2}}}{3} + \frac{2}{3}e^{\frac{y_{2}}{2}}\cos\left( \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{- y_{2}}}{3} - \frac{2}{3}e^{\frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{2}}{2} \right) \\ \frac{e^{- y_{2}}}{3} - \frac{2}{3}e^{\frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{{- y}_{2}}}{3} - \frac{2}{3}e^{\frac{y_{2}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{2}}{2} \right) & \frac{e^{{- y}_{2}}}{3} + \frac{2}{3}e^{\frac{y_{2}}{2}}\cos\left( \frac{\sqrt{3}y_{2}}{2} \right) \end{pmatrix} \end{array} \right\}(1.1.110) \end{array}
GY3:{GY3(y3)=(ey33+23e−y32cos⁡(3y32)ey33−23e−y32sin⁡(π6−3y32)ey33−23e−y32sin⁡(π6+3y32)ey33−23e−y32sin⁡(π6+3y32)ey33+23e−y32cos⁡(3y32)ey33−23e−y32sin⁡(π6−3y32)ey33−23e−y32sin⁡(π6−3y32)ey33−23e−y32sin⁡(π6+3y32)ey33+23e−y32cos⁡(3y32)),GY3(−y3)=(e−y33+23ey32cos⁡(3y32)e−y33−23ey32sin⁡(π6+3y32)e−y33−23ey32sin⁡(π6−3y32)e−y33−23ey32sin⁡(π6−3y32)e−y33+23ey32cos⁡(3y32)e−y33−23ey32sin⁡(π6+3y32)e−y33−23ey32sin⁡(π6+3y32)e−y33−23ey32sin⁡(π6−3y32)e−y33+23ey32cos⁡(3y32))}(1.1.111)G_{Y_{3}}:\left\{ \begin{array}{r} G_{Y_{3}}\left( y_{3} \right) = \begin{pmatrix} \frac{e^{y_{3}}}{3} + \frac{2}{3}e^{- \frac{y_{3}}{2}}\cos\left( \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{y_{3}}}{3} - \frac{2}{3}e^{- \frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{y_{3}}}{3} - \frac{2}{3}e^{- \frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{3}}{2} \right) \\ \frac{e^{y_{3}}}{3} - \frac{2}{3}e^{- \frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{y_{3}}}{3} + \frac{2}{3}e^{- \frac{y_{3}}{2}}\cos\left( \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{y_{3}}}{3} - \frac{2}{3}e^{- \frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{3}}{2} \right) \\ \frac{e^{y_{3}}}{3} - \frac{2}{3}e^{- \frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{y_{3}}}{3} - \frac{2}{3}e^{- \frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{y_{3}}}{3} + \frac{2}{3}e^{- \frac{y_{3}}{2}}\cos\left( \frac{\sqrt{3}y_{3}}{2} \right) \end{pmatrix}, \\ G_{Y_{3}}\left( {- y}_{3} \right) = \begin{pmatrix} \frac{e^{{- y}_{3}}}{3} + \frac{2}{3}e^{\frac{y_{3}}{2}}\cos\left( \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{- y_{3}}}{3} - \frac{2}{3}e^{\frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{{- y}_{3}}}{3} - \frac{2}{3}e^{\frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{3}}{2} \right) \\ \frac{e^{{- y}_{3}}}{3} - \frac{2}{3}e^{\frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{{- y}_{3}}}{3} + \frac{2}{3}e^{\frac{y_{3}}{2}}\cos\left( \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{- y_{3}}}{3} - \frac{2}{3}e^{\frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{3}}{2} \right) \\ \frac{e^{- y_{3}}}{3} - \frac{2}{3}e^{\frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} + \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{{- y}_{3}}}{3} - \frac{2}{3}e^{\frac{y_{3}}{2}}\sin\left( \frac{\pi}{6} - \frac{\sqrt{3}y_{3}}{2} \right) & \frac{e^{{- y}_{3}}}{3} + \frac{2}{3}e^{\frac{y_{3}}{2}}\cos\left( \frac{\sqrt{3}y_{3}}{2} \right) \end{pmatrix} \end{array} \right\}(1.1.111)

We can see the re­sem­blance the GXjG_{X_{j}} ro­ta­tion ma­tri­ces have with the Rx, Ry, RzR_{x},\ R_{y},\ R_{z} ma­tri­ces from the SO(3)SO(3) group, where the latter group has the fol­low­ing gen­er­a­tors

Gx=(00000−1010),  Gy=(001000−100),  Gz=(0−10100000).\begin{array}{r} G_{x} = \begin{pmatrix} 0 & 0 & 0 \\ 0 & 0 & - 1 \\ 0 & 1 & 0 \end{pmatrix},\ \ G_{y} = \begin{pmatrix} 0 & 0 & 1 \\ 0 & 0 & 0 \\ - 1 & 0 & 0 \end{pmatrix},\ \ G_{z} = \begin{pmatrix} 0 & - 1 & 0 \\ 1 & 0 & 0 \\ 0 & 0 & 0 \end{pmatrix}. \end{array}
(1.1.112)

An el­e­ment of SO(3)SO(3) would be a linear com­bi­na­tion of the above gen­er­a­tors

α1Gx+α2Gy+α3Gz∈SO(3),   αj∈ℝ3.\begin{array}{r} \alpha_{1}G_{x} + \alpha_{2}G_{y} + \alpha_{3}G_{z} \in SO(3),\ \ \ \alpha_{j} \in \mathbb{R}^{3}. \end{array}
(1.1.113)

Ex­po­nen­ti­a­tion of the GjG_{j} gen­er­a­tors yield the ma­tri­ces

Rx=(1000cos⁡(α)−sin⁡(α)0sin⁡(α)cos⁡(α)),   Ry=(cos⁡(α)0sin⁡(α)010)−sin⁡(α)0cos⁡(α)),   Rz=(cos⁡(α)−sin⁡(α)0sin⁡(α)cos⁡(α)0001).\begin{array}{r} R_{x} = \begin{pmatrix} 1 & 0 & 0 \\ 0 & \cos(\alpha) & - \sin(\alpha) \\ 0 & \sin(\alpha) & \cos(\alpha) \end{pmatrix},\ \ \ R_{y} = \begin{pmatrix} \cos(\alpha) & 0 & \sin(\alpha) \\ 0 & 1 & 0) \\ - \sin(\alpha) & 0 & \cos(\alpha) \end{pmatrix},\ \ \ R_{z} = \begin{pmatrix} \cos(\alpha) & - \sin(\alpha) & 0 \\ \sin(\alpha) & \cos(\alpha) & 0 \\ 0 & 0 & 1 \end{pmatrix}. \end{array}
(1.1.114)

With our U(3)U(3) group, in­stead of ro­ta­tions in ℝ3\mathbb{R}^{3}, we have ro­ta­tions that are a linear com­bi­na­tion of three com­plex planes in ℂ3\mathbb{C}^{3}. We use the ex­am­ple of our X1X_{1} matrix below to com­pare with the Rx∈SO(3)R_{x} \in SO(3) matrix above, show­ing the com­plex axis of ro­ta­tion as 𝐞𝐢𝐗𝟏\mathbf{e}^{\mathbf{i}\mathbf{X}_{\mathbf{1}}}, with the other en­tries rep­re­sent­ing a ro­ta­tion in the plane of X2X3∈ℂ2X_{2}X_{3} \in \mathbb{C}^{2}

eX1=(𝐞𝐢𝐗𝟏000cos⁡(X1)isin(X1)0isin(X1)cos⁡(X1)).\begin{array}{r} e^{X_{1}} = \begin{pmatrix} \mathbf{e}^{\mathbf{i}\mathbf{X}_{\mathbf{1}}} & 0 & 0 \\ 0 & \cos\left( X_{1} \right) & isin\left( X_{1} \right) \\ 0 & isin\left( X_{1} \right) & \cos\left( X_{1} \right) \end{pmatrix}. \end{array}
(1.1.115)

By taking the com­plex mod­u­lus of 𝐞𝐢𝐗𝟏\mathbf{e}^{\mathbf{i}\mathbf{X}_{\mathbf{1}}}, and using the iso­mor­phism i≅(0∓1±10)i \cong \begin{pmatrix} 0 & \mp 1 \\ \pm 1 & 0 \end{pmatrix}, ℝ3\mathbb{R}^{3} could be seen as a pro­jec­tive form of ℂ3\mathbb{C}^{3},

eX1=(|𝐞𝐢𝐗𝟏|000cos⁡(X1)isin(X1)0isin(X1)cos⁡(X1))ℂ3→ℝ3→(1000cos⁡(X1)∓sin⁡(X1)0±sin⁡(X1)cos⁡(X1)).\begin{array}{r} e^{X_{1}} = \begin{pmatrix} \left| \mathbf{e}^{\mathbf{i}\mathbf{X}_{\mathbf{1}}} \right| & 0 & 0 \\ 0 & \cos\left( X_{1} \right) & isin\left( X_{1} \right) \\ 0 & isin\left( X_{1} \right) & \cos\left( X_{1} \right) \end{pmatrix}\mathbb{C}^{3} \rightarrow \mathbb{R}^{3} \rightarrow \begin{pmatrix} 1 & 0 & 0 \\ 0 & \cos\left( X_{1} \right) & \mp \sin\left( X_{1} \right) \\ 0 & {\pm \sin}\left( X_{1} \right) & \cos\left( X_{1} \right) \end{pmatrix}. \end{array}
(1.1.116)

The sim­i­lar­i­ties be­tween our ex­po­nen­ti­at­ed XjX_{j} gen­er­a­tors and that of SO(3)SO(3)) begets the ques­tion, is the three-di­men­sion­al space we are all fa­mil­iar with ac­tu­al­ly ℂ3\mathbb{C}^{3}? We would argue that it is indeed. The notion of need­ing a real-axis to rotate about in space is not nec­es­sary. To show this, we call at­ten­tion to U(1)U(1), and its real form SO(2)SO(2), which have no fixed real axes but rotate in two-di­men­sion­al space seam­less­ly. In the U(3)U(3) group, the XjX_{j} gen­er­a­tor ro­ta­tions ∈ℂ3\in \mathbb{C}^{3} do not change the com­plex plane that the axis func­tion of the XjX_{j} group sits in. This occurs while an inter-planar ro­ta­tion occurs with the ar­gu­ment for XjX_{j} in the XkXlX_{k}X_{l} sub­space. Look­ing at Eq. 1.1.116 above, 𝐞𝐢𝐗𝟏\mathbf{e}^{\mathbf{i}\mathbf{X}_{\mathbf{1}}} is the fixed two-di­men­sion­al "axis" ∈ℂ,\in \mathbb{C},while the pri­ma­ry ro­ta­tion occurs in the ℂ2\mathbb{C}^{2}-plane, X2/X3.X_{2}{/X}_{3}.The rep­re­sen­ta­tion in the polar form of the whole group can be cal­cu­lat­ed by taking ad­van­tage of the fact that the al­ge­bra 𝔤\mathfrak{g} is di­ag­o­nal­iz­able

𝔤=VDV−1.\begin{array}{r} \mathfrak{g} =VDV^{- 1}. \end{array}
(1.1.117)

The al­ge­bra set 𝔤={Xj,Yj}\mathfrak{g} =\left\{ X_{j},Y_{j} \right\} can be put into matrix form as

𝔤={(y1000y1000y1)(x1y2+x3y3+x2y3+x3x2y2+x1y2+x2y3+x1x3)}.\begin{array}{r} \mathfrak{g} =\left\{ \begin{pmatrix} y_{1} & 0 & 0 \\ 0 & y_{1} & 0 \\ 0 & 0 & y_{1} \end{pmatrix}\begin{pmatrix} x_{1} & {y_{2} + x}_{3} & y_{3} + x_{2} \\ {y_{3} + x}_{3} & x_{2} & {y_{2} + x}_{1} \\ {y_{2} + x}_{2} & {y_{3} + x}_{1} & x_{3} \end{pmatrix} \right\}. \end{array}
(1.1.118)

In the eigen­de­com­po­si­tion 𝔤=Y1(VDV−1)\mathfrak{g} =Y_{1}\left( VDV^{- 1} \right), the group iden­ti­ty Y1Y_{1} will simply scale the VDV−1VDV^{- 1} ma­tri­ces uni­form­ly by e1e^{1}, through the ex­po­nen­tial map­ping. The eigen­val­ue and eigen­vec­tor ma­tri­ces D,VD,V have the fol­low­ing repere­sen­ta­tions

D=(−(ℛ2)−𝛌2000−(ℛ2)+𝛌𝟐000 ℛ+𝛄),\begin{array}{r} D = \begin{pmatrix} - \left( \frac{\mathcal{R}}{2} \right) - \frac{\mathbf{\lambda}}{2} & 0 & 0 \\ 0 & - \left( \frac{\mathcal{R}}{2} \right) + \frac{\mathbf{\lambda}}{\mathbf{2}} & 0 \\ 0 & 0 & \ \mathcal{R} +\mathbf{\gamma} \end{pmatrix}, \end{array}
(1.1.119)
V=((y2−y3−2x1+2x3+𝛌2(x1−x2−y2+y3))−(y2−y3−2x1+2x3+𝛌2(x1−x2−y2+y3))1−(y2−y3−2x1+2x3+𝛌2(x1−x2−y2+y3))(y2−y3−2x1+2x3+𝛌2(x1−x2−y2+y3))111 1),\begin{array}{r} V = \begin{pmatrix} \left( \frac{y_{2} - y_{3} - 2x_{1} + 2x_{3} + \mathbf{\lambda}}{2\left( x_{1} - x_{2} - y_{2} + y_{3} \right)} \right) & - \left( \frac{y_{2} - y_{3} - 2x_{1} + 2x_{3} + \mathbf{\lambda}}{2\left( x_{1} - x_{2} - y_{2} + y_{3} \right)} \right) & 1 \\ - \left( \frac{y_{2} - y_{3} - 2x_{1} + 2x_{3} + \mathbf{\lambda}}{2\left( x_{1} - x_{2} - y_{2} + y_{3} \right)} \right) & \left( \frac{y_{2} - y_{3} - 2x_{1} + 2x_{3} + \mathbf{\lambda}}{2\left( x_{1} - x_{2} - y_{2} + y_{3} \right)} \right) & 1 \\ 1 & 1 & \ 1 \end{pmatrix}, \end{array}
(1.1.120)

for ℛ=(y2+y3)\mathcal{R} =\left( y_{2} + y_{3} \right),   γ=(x1+x2+x3),  𝛌= −3y22+6y2y3−3y32+4x12−4x1x2−4x1x3+4x22−4x2x3+4x32.\ \ \gamma = \left( x_{1} + x_{2} + x_{3} \right),\ \ \mathbf{\lambda} = \ \sqrt{\begin{array}{r} - 3y_{2}^{2} + 6y_{2}y_{3} - 3y_{3}^{2} + 4x_{1}^{2} \\ - 4x_{1}x_{2} - 4x_{1}x_{3} + 4x_{2}^{2} - 4x_{2}x_{3} + 4x_{3}^{2} \end{array}}.

We can take ad­van­tage of the di­ag­o­nal­ized form to cal­cu­late the polar form of the entire group

G=VeDV−1,\begin{array}{r} G = Ve^{D}V^{- 1}, \end{array}
(1.1.121)

find­ing the matrix rep­re­sen­ta­tion G=ey1(VeDV−1)G = e^{y_{1}}\left( Ve^{D}V^{- 1} \right), as

ey1(e(γ+ℛ)3+e−ℛ2(2cosh⁡(𝛌2)3−i∂𝛌∂X1sinh⁡(𝛌2)3)e(γ+ℛ)3−e−ℛ2(cosh⁡(𝛌2)3+∂𝛌∂Y2+i∂𝛌∂X3sinh⁡(𝛌2)3)e(γ+ℛ)3−e−ℛ2(cosh⁡(𝛌2)3+−∂𝛌∂Y3−i∂𝛌∂X2sinh⁡(𝛌2)3)e(γ+ℛ)3−e−ℛ2(cosh⁡(𝛌2)3+−∂𝛌∂Y3−i∂𝛌∂X3sinh⁡(𝛌2)3)e(γ+ℛ)3+e−ℛ2(2cosh⁡(𝛌2)3−i∂𝛌∂X2sinh⁡(𝛌2)3)e(γ+ℛ)3−e−ℛ2(cosh⁡(𝛌2)3+∂𝛌∂Y2+i∂𝛌∂X1sinh⁡(𝛌2)3)e(γ+ℛ)3−e−ℛ2(cosh⁡(𝛌2)3+∂𝛌∂Y2+i∂𝛌∂X2sinh⁡(𝛌2)3)e(γ+ℛ)3−e−ℛ2(cosh⁡(𝛌2)3+−∂𝛌∂Y3−i∂𝛌∂X1sinh⁡(𝛌2)3)e(γ+ℛ)3+e−ℛ2(2cosh⁡(𝛌2)3−i∂𝛌∂X3sinh⁡(𝛌2)3))\begin{array}{r} e^{y_{1}}\begin{pmatrix} \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} + e^{- \frac{\mathcal{R}}{2}}\left( \frac{2\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} - \frac{i\frac{\partial\mathbf{\lambda}}{\partial X_{1}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) & \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} - e^{- \frac{\mathcal{R}}{2}}\left( \frac{\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} + \frac{\frac{\partial\mathbf{\lambda}}{\partial Y_{2}} + i\frac{\partial\mathbf{\lambda}}{\partial X_{3}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) & \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} - e^{- \frac{\mathcal{R}}{2}}\left( \frac{\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} + \frac{- \frac{\partial\mathbf{\lambda}}{\partial Y_{3}} - i\frac{\partial\mathbf{\lambda}}{\partial X_{2}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) \\ \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} - e^{- \frac{\mathcal{R}}{2}}\left( \frac{\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} + \frac{- \frac{\partial\mathbf{\lambda}}{\partial Y_{3}} - i\frac{\partial\mathbf{\lambda}}{\partial X_{3}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) & \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} + e^{- \frac{\mathcal{R}}{2}}\left( \frac{2\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} - \frac{i\frac{\partial\mathbf{\lambda}}{\partial X_{2}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) & \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} - e^{- \frac{\mathcal{R}}{2}}\left( \frac{\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} + \frac{\frac{\partial\mathbf{\lambda}}{\partial Y_{2}} + i\frac{\partial\mathbf{\lambda}}{\partial X_{1}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) \\ \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} - e^{- \frac{\mathcal{R}}{2}}\left( \frac{\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} + \frac{\frac{\partial\mathbf{\lambda}}{\partial Y_{2}} + i\frac{\partial\mathbf{\lambda}}{\partial X_{2}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) & \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} - e^{- \frac{\mathcal{R}}{2}}\left( \frac{\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} + \frac{- \frac{\partial\mathbf{\lambda}}{\partial Y_{3}} - i\frac{\partial\mathbf{\lambda}}{\partial X_{1}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) & \frac{e^{\left( \gamma+ \mathcal{R} \right)}}{3} + e^{- \frac{\mathcal{R}}{2}}\left( \frac{2\cosh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} - \frac{i\frac{\partial\mathbf{\lambda}}{\partial X_{3}}\sinh\left( \frac{\mathbf{\lambda}}{2} \right)}{3} \right) \end{pmatrix} \end{array}
Figure 1.1.11 – The Com­plete G=ey1(VeDV−1)G = e^{y_{1}}\left( Ve^{D}V^{- 1} \right) Lie Group Rep­re­sen­ta­tion

for par­tial de­riv­a­tives ∂𝛌∂𝐗𝐣\frac{\partial\mathbf{\lambda}}{\partial\mathbf{X}_{\mathbf{j}}}, rep­re­sent­ed by

∂𝛌∂𝐗𝟏=(2x1−x2−x3)𝛌,   ∂𝛌∂𝐗𝟐=(−x1+2x2−x3)𝛌,   ∂𝛌 ∂𝐗𝟑=(−x1−x2+2x3)𝛌.\begin{array}{r} \frac{\partial\mathbf{\lambda}}{\partial\mathbf{X}_{\mathbf{1}}} = \frac{\left( 2x_{1} - x_{2} - x_{3} \right)}{\mathbf{\lambda}},\ \ \ \frac{\partial\mathbf{\lambda}}{\partial\mathbf{X}_{\mathbf{2}}} = \frac{\left( - x_{1} + 2x_{2} - x_{3} \right)}{\mathbf{\lambda}},\ \ \ \frac{\partial\mathbf{\lambda}}{\ \partial\mathbf{X}_{\mathbf{3}}} = \frac{\left( - x_{1} - x_{2} + 2x_{3} \right)}{\mathbf{\lambda}}. \end{array}

It is also noted that for the polar form of the group G,G,we have

G∗G=I,\begin{array}{r} G^{*}G = I, \end{array}
(1.1.122)

and

det⁡(G)=e(ix1+ix2+ix3)∴|det⁡(G)| =I.\begin{array}{r} \det(G) = e^{\left( ix_{1} + ix_{2} + ix_{3} \right)}\therefore|\det(G)|\ = I. \end{array}
(1.1.123)

In the U(3)U(3) group, we have ex­tend­ed the sym­me­try group S3S_{3} to the com­plex field of ℂ3\mathbb{C}^{3}. The group's bound­ary, by its uni­tary char­ac­ter­is­tic, forms a sphere ∈ℂ3\in \mathbb{C}^{3} with a uni­form mod­u­lus of one. It has six di­men­sions, three real and three imag­i­nary, which fits the sym­me­try a per­fect sphere might embody.

  1. Hall, Brian C. – "Lie Groups, Lie Al­ge­bras, and Rep­re­sen­ta­tions", Grad­u­ate Texts in Math­e­mat­ics, Springer, 2015