1:
2:
3:
4: %\documentstyle[aps,preprint]{revtex}
5: %\newcommand{\be}{\begin{eqnarray}}
6: %\newcommand{\e}{\end{eqnarray}}
7: %\newcommand{\pp}{\partial}
8: %\setcounter{page}{1}
9: %\begin{document}
10: %\newcommand{\kw}{\kappa}
11: %\setcounter{section}{0}
12: %\renewcommand{\thesection}{1.\arabic{section}}
13: %\setcounter{equation}{0}
14: %\renewcommand{\theequation}{1.\arabic{equation}}
15: %{\flushleft\huge\bf {Chapter 1}}
16: \chapter{Introduction}
17: %\vskip .2in
18: %%%%%%%%%%%%%%%%%%%%%%%%%%%
19: \section{Motivation}\label{mot}
20: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
21: Lepton-hadron deep inelastic scattering (DIS) is a very important tool to
22: unravel the structure of
23: hadrons. To the lowest order, this scattering takes place via the exchange
24: of a virtual photon.
25: When a very low mass virtual photon ($Q^2<1 Gev^2$) scatters off a proton,
26: the photon `sees' only the total charge and magnetic moment of the proton
27: and the proton appears to be a pointlike object. However, a higher virtual mass photon
28: resolves the fine structure of the proton charge distributions and sees its
29: elementary constituents.
30: The cross section of such scattering can be expressed in terms of several
31: functions which depend on two variables, $Q^2$ (momentum transfer square
32: in the process) and $\nu$ (energy of the virtual photon). These functions, to
33: start with, are unknown and they depend on the hadron structure. They are
34: called structure functions.
35: Early SLAC experimental results showed that the
36: structure functions show a phenomenon called scaling, which means that they
37: depend only on the ratio of $Q^2$ and $\nu$, and do not depend
38: on them separately. In the pre-QCD era, Bjorken used current algebra
39: techniques to explain such behavior of structure functions. Later, Feynman
40: proposed parton model idea according to which, the proton consists of
41: pointlike, massless, noninteracting particles called partons and the
42: scattering takes place incoherently from the partons. The parton model
43: successfully explained scaling of the structure functions. However,
44: substantial scaling violations were observed in later SLAC experiments and
45: the structure functions were found to evolve with $Q^2$. This violation of
46: scaling occurs because of the interactions between the constituents of the
47: proton, which are now known as the quarks and gluons. Their interaction is
48: governed by QCD (quantum chromodynamics).
49:
50:
51: The $Q^2$ dependence of the structure functions can be logarithmic or there
52: may be power dependence. The contributions to the structure functions which
53: are proportional to some power of ${1\over Q}$ are commonly called `higher
54: twist' contributions. The most widely used formal approach to DIS structure
55: functions is based on Wilson's operator product expansion (OPE). There
56: exists a more formal definition of twist in terms of OPE, which we shall
57: discuss shortly. DIS is a high energy process, and $Q^2$ is very large. This
58: means that higher and higher twist contributions are suppressed compared to
59: the leading twist contributions. However, recent experiments indicate that
60: the higher twist effects play a very important role in the SLAC kinematical
61: range. Understanding the higher twist effects requires non-trivial
62: non-perturbative information of the structure of the hadron. So far, a clear
63: and physical picture of the higher twist contribution is lacking. To
64: understand these effects, there is an urgent need to develop a
65: non-perturbative technique to analyze DIS. For this, one needs theoretical
66: tools which are based on physical intuitions and at the same time employs
67: well-defined field theoretic calculational procedure.
68:
69: The usual field theoretical formulation is Lagrangian formulation, in which
70: one starts from the local Lagrangian, gets the equations of motion and
71: then quantizes the system by imposing suitable commutation or
72: anti-commutation relations between the fields and their conjugate momenta on
73: a space-like hypersurface. It is a manifestly covariant formalism. However,
74: for interacting theory, the equations of motion involve non-trivial
75: interacting fields and can only be solved approximately, using
76: perturbative techniques. Perturbative methods cannot be applied
77: to low energy QCD bound state problems since the coupling constant takes a
78: large value in this energy range. Another point is that, in the covariant
79: Lagrangian formulation, the intuitive picture of quantum mechanics is lost.
80: There exists another formulation known as Hamiltonian formulation of quantum field theory
81: in which one starts with the Hamiltonian, obtains the Hamiltonian equations
82: of motion and sets the quantization conditions on a space-like hypersurface.
83: The Hamiltonian gives the time evolution of the system. Though it is physically
84: more intuitive, there are some disadvantages of Hamiltonian formulation. It
85: is manifestly non-covariant since one has to choose a time axis and perturbation theory becomes more complicated.
86: The latter is due to the fact that here one has to do old-fashioned time
87: ordered perturbation theory and instead of Feynman diagrams, one has to
88: calculate a larger number of time-ordered diagrams at each order. It has
89: been shown that, light-front quantization (where the surface of quantization
90: is tangential to the light cone) makes Hamiltonian formulation simpler. This
91: is due to the fact that. the diagrams which produce particles from the vacuum
92: are absent here (this point will be clarified in chapter 3). So one
93: encounters lesser number of time ordered diagrams. In
94: what follows, we have used Hamiltonian light-front QCD framework
95: (Hamiltonian formulation of QCD where the quantization surface is tangential
96: to the light-cone) to analyze DIS structure functions.
97:
98: In order to get an intuitive picture of DIS, it is necessary to keep close
99: contact with Feynman's parton model. As mentioned before, partons were
100: originally introduced as collinear, massless, non-interacting constituents
101: of the proton. It is necessary to generalize the parton ideas to introduce
102: field theoretic partons which are massive, non-collinear and their
103: interaction is governed by QCD. This goal is achieved in light-front
104: Hamiltonian QCD analysis.
105: This is
106: essentially a non-perturbative method which allows us to explore the
107: non-perturbative higher twist effects explicitly. In this approach, instead
108: of calculating the moments of the structure functions as one does in the
109: conventional OPE formalism, one directly calculates the structure functions
110: themselves which involve hadron matrix elements of light-front bilocal operators.
111: The structure functions can be calculated once the light-front wave
112: functions are known. However, perturbative calculations are also possible,
113: and both perturbative and non-perturbative aspects can be
114: investigated within the same framework.
115:
116: In what follows, we have used light-front Hamiltonian QCD framework
117: to study the twist four part of the longitudinal
118: structure function $F_L$ in unpolarized DIS and the transverse polarized structure
119: function $g_T$ in polarized DIS, which is a twist three contribution. Both
120: of these structure functions involve non-trivial operators which brings in
121: non-perturbative dynamics. It is well known that the evolution of $F_L$ and
122: $g_T$ with the scale, $Q^2$, are highly complicated. We have shown that the reason of this
123: complexity is the fact that $F_L$ is related to the light-front QCD
124: Hamiltonian density and $g_T$ is related to the transverse spin operator,
125: both of which are complicated dynamical operators (they change the surface
126: of quantization, see chapter 3). As a result, both these structure
127: functions acquire off-diagonal contributions from the matrix elements, in
128: contrast to the other two electroproduction structure functions, $F_2$ and
129: $g_1$, which involve diagonal matrix elements only. Another interesting
130: point is, the operators involved acquire divergences in perturbation theory.
131: The experimentally measured structure functions, of course, are finite
132: quantities. Therefore, one must know how to remove these divergences
133: (renormalization). We have addressed the issue of renormalization of these
134: composite operators which, in light-front theory is highly non-trivial. As
135: we mentioned before, a very interesting point is that, both perturbative and
136: non-perturbative issues can be address in our approach. In
137: the non-perturbative context, we have analyzed twist-four longitudinal
138: structure function for a meson in $1+1$ dimensional QCD and for a
139: positronium in $3+1$ dimensional QED in weak coupling limit.
140:
141: Earlier, it has been shown that the integral of the structure function
142: $F_2$ is related to the light-front longitudinal momentum density and the
143: integral of the polarized structure function $g_1$ is related to the
144: helicity. In this work, we have shown that the integral of $F_L$ is directly related to the
145: light-front Hamiltonian density. Also the integral of the transverse
146: polarized structure function $g_T$ is related to the transverse spin
147: operator in light-front QCD. Thus our approach gives a unified picture of
148: DIS structure functions and connects them to the light-front Poincare
149: generators. We emphasize that such connections arise because our approach
150: is different from the usual one and we get many new and interesting results.
151: \vskip .2in
152: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
153: \section{Plan of the Thesis}\label{plan}
154: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
155: The usual approach to analyze DIS structure functions is based on OPE. This
156: mainly concentrates on the
157: evolution of the structure functions in perturbative QCD. In these methods,
158: the moments of the structure functions automatically arise. We briefly
159: discuss the OPE method to analyze DIS structure functions in chapter 2. In
160: this chapter, we also explain the term `twist' in the context of OPE.
161: It is also important to understand the origin and nature of the higher twist
162: contributions. We discuss how these higher twist effects are
163: addressed in the covariant formalism. For the sake of simplicity, we briefly
164: outline the procedure and omit the mathematical details.
165: The effect of the target mass gives a higher twist contribution, which, in
166: the free theory assumption produces $\xi$ scaling. The dynamical higher
167: twist effects are highly complicated. We point out that
168: in the QCD improved parton model approach, these contributions are expressed
169: as the Fourier transform of the target matrix elements of various
170: bilocal operators. However, these matrix elements cannot be further
171: investigated because of the non-perturbative hadronic state.
172:
173: We follow an alternative approach based on light-front
174: Hamiltonian QCD. The light-front formulation of quantum field theory has
175: many strikingly different features compared to the usual equal time
176: formulation. Some of these features makes it more suitable for the analysis
177: of DIS process. It is important to remember that the parton model was
178: formulated in the infinite momentum frame, which, in some sense is
179: equivalent to light-front formulation. In chapter 3, we discuss the basic
180: features of light-front field theory and light-front QCD.
181:
182: In chapter 4, we outline our approach. Bjorken-Johnson-Low (high energy)
183: expansion of the Compton scattering amplitude gives the expressions of the
184: structure functions as the Fourier transform of the matrix elements of
185: light-front bilocal operators. We work in the light-front gauge. In this method, one gets the structure
186: functions themselves and not their moments. By expanding the target state in
187: Fock space and using truncation, the structure functions are expressed in
188: terms of light cone wave functions. Splitting functions
189: are obtained in perturbation theory simply by replacing the target by a
190: dressed parton. In this chapter, we review the works done so far in
191: this approach. We show that this method not only reproduces the well known results of the
192: covariant formalism, like the $Q^2$ evolution of the twist two structure
193: function $F_2$, but also addresses some important issues, like the effect of
194: quark mass in $g_T$ and the helicity sum rule of the proton, in a more
195: transparent way.
196:
197: The actual work done in this research program has been discussed in detail
198: in chapters 5, 6, 7 and 8. These works are concerned with the higher twist
199: effects $F_L$ and $g_T$.
200: All of these results are new and interesting, both from theoretical and
201: phenomenological point of view. Since our approach is different from the
202: usual one, such new outcomes are not unexpected. Also, our approach gives
203: better insights into problems that are not easy to understand in the
204: commonly used procedure.
205:
206: In chapter 5, we discuss a new sum rule that we
207: have obtained which connects the twist four part of $F_L$ to the fermionic
208: part of light-front QCD Hamiltonian density. Introducing a
209: gluonic part of $F_L$, we show that the sum rule relates the total $F_L$ to
210: the target mass. To our knowledge, this is the first sum rule for a higher
211: twist structure function. In this chapter, we also perform one loop renormalization
212: of $F_L$ in light-front Hamiltonian perturbation theory. Replacing the
213: target by a dressed quark, we show that the twist four part of $F_L$ is
214: directly related to the quark mass shift in perturbation theory.
215:
216:
217: In chapter 6, we continue our investigation of $F_L$. In the non-perturbative
218: context, we verify that the sum rule is obeyed for a meson in $1+1$ dimensional
219: QCD because of t' Hooft equation. An analysis for a bound state in QCD in $3+1$
220: dimension is very complicated and there one has to use the recently
221: developed similarity renormalization technique. In order to understand the
222: calculational procedure, we consider a simpler case, namely a
223: positronium-like bound state in $3+1$ dimensional QED in the weak coupling
224: limit. We explicitly verify the sum rule in this case.
225:
226: In chapters 7 and 8, we investigate another most challenging problem,
227: namely that of the spin operator for a composite system in an arbitrary
228: reference frame in relativistic quantum field theory. We
229: show that, this issue is very important in the context of polarized
230: scattering, since the transverse polarized structure function $g_T$ is
231: related to a part of the transverse spin operator. It is also known that
232: $g_1$ is related to the helicity operator. In chapter 7, we
233: discuss the problems associated with the definitions of the relativistic
234: spin operators in equal-time theory and contrast with the light-front
235: case. The light-front theory is more suitable for this purpose because boost
236: is kinematical.
237: We discuss the light-front spin operators for the massive and
238: massless cases separately. The transverse spin operator contains the
239: transverse rotation operator which is interaction dependent. We
240: construct the transverse rotation operator in light-front QCD. In this
241: chapter, we also show that though the light-front transverse spin
242: operators cannot be separated into orbital and spin parts, there exists a
243: decomposition of it which is physically interesting.
244:
245: In chapter 8, We explore the physical significances of this
246: decomposition and compare and contrast it with the decomposition of the
247: helicity operator into orbital and spin parts. We also perform one
248: loop renormalization of the full transverse spin operator. Our results show
249: that only one counterterm
250: is needed and that has the same form as the linear mass counterterm in light-front
251: Hamiltonian perturbation theory. We explicitly show that in an arbitrary
252: reference frame, all the center of mass momentum dependence get canceled in
253: the matrix element of the transverse spin operator. We emphasize that in the
254: equal time case, this cancellation is almost impossible to prove because of
255: the dynamical boost operators.
256:
257: Summary and conclusions are given in chapter 9. References are given at the
258: end of each chapter. In order to clarify the
259: notations and also the intermediate steps of calculations, we have provided
260: several appendices.
261: