betaFeedback

WorksPhysics and Astronomy

Modified Homotopic approach for diffractive production

Abstract

We review the recent developments of the use of the homotopy method for solving the non-linear evolution equation for the diffractive production in deep inelastic scattering. We introduce part of the non-linear corrections in the linear term. This simplified non-linear evolution equation is solved analytically taking into account the initial and boundary conditions for the process. It turns out that these corrections are rather small and can be estimated in the regular iterative procedure.

Introduction

In this paper we present a procedure to solve the non-linear equations for diffractive process that in QCD govern the dynamics in the saturation region. In our previous paper [1, 2] we found a solution of the Balitsky-Kovchegov (BK) equation [3, 4] that gives the dipole scattering amplitude using the homotopy approach. It has been shown that this approach allows us to collect all essential contribution into the linear equation which can be solved analytically, and to propose an iteration procedure, which is partly numerical and leads to small corrections.

For diffractive production we have several different kinematic regions for NDN^D, where σdipolediff(r⊥,Y,Y0)=∫d2b ND(r⊥,Y,Y0;b)\sigma_{\mathrm{dipole}}^{\mathrm{diff}}(r_{\perp},Y,Y_{0}) = \int d^{2}b\,N^D(r_{\perp},Y,Y_{0};b) is the cross section of diffractive production with the rapidity gap larger than Y0Y_0. The non-linear evolution equation for ND(Y,Y0,r10;b)N^D(Y,Y_0,r_{10};b) that describes these diffraction production in deep inelastic scattering has been derived in Ref. [5] and has the following form:

∂ND(Y,Y0,r10;b)∂Y=αˉs2π∫d2r2r012r022r122{N12D+N20D−N10D+N12DN20D−2N12DN20−2N12N20D+2N12N20},(1)\frac{\partial N^D(Y,Y_0,r_{10};b)}{\partial Y} = \frac{\bar{\alpha}_s}{2\pi} \int d^2r_2 \frac{r_{01}^2}{r_{02}^2r_{12}^2} \left\{N^D_{12} + N^D_{20} - N^D_{10} + N^D_{12}N^D_{20} - 2N^D_{12}N_{20} - 2N_{12}N^D_{20} + 2N_{12}N_{20}\right\}, \tag*{(1)}

where Nik=N(Y;rik;b)N_{ik} = N(Y;r_{ik};b) is the elastic scattering amplitude of a dipole with size rikr_{ik} and rapidity YY and NikD=ND(Y;Y0,rik;b)N^D_{ik} = N^D(Y;Y_0,r_{ik};b) is the cross section of the diffractive production with the rapidity gap larger than Y0Y_0 at the impact parameter bb for the same scattering process.

In general, the homotopy method can be used as an effective procedure for solving an equation of the form L[u]+NL[u]=0\mathcal{L}[u] + \mathcal{N}_{\mathcal{L}}[u] = 0, where the linear part L[u]\mathcal{L}[u] is a differential or integral-differential operator, and the non-linear part NL[u]\mathcal{N}_{\mathcal{L}}[u] has an arbitrary form. To solve it, we introduce the following equation for the homotopy function H(p,u)\mathcal{H}(p,u): H(p,u)=L[up]+pNL[up]=0\mathcal{H}(p,u) = \mathcal{L}[u_p] + p\mathcal{N}_{\mathcal{L}}[u_p] = 0. Solving this expression, we reconstruct the function

up(Y,x10,b)=u0(Y,x10,b)+pu1(Y,x10,b)+p2u2(Y,x10,b)+…(2)u_p(Y,x_{10},b) = u_0(Y,x_{10},b) + p u_1(Y,x_{10},b) + p^2 u_2(Y,x_{10},b) + \ldots \tag*{(2)}

with L[u0]=0\mathcal{L}[u_0] = 0. (2) gives the solution to the non-linear equation for p=1p = 1. The hope is that several terms in this series will give a good approximation to the solution of the non-linear equation.

The linear equation is obvious in the perturbative QCD region (see Figure 1), where it is the BFKL equation. However, we will show that inside the saturation region (see Figure 1) we can find the linearized equation based on the approach of Ref. [6]. We demonstrate that the non-linear term, which includes the remains of the non-linear corrections, leads to small contributions and can be treated in the perturbation approach. This aspect has been discussed in Ref. [2], whose results are here briefly described.

Figure 1. Saturation region of QCD for the elastic scattering amplitude with κ=χ(γcr)1−γcr\kappa= \frac{\chi(\gamma_{\mathrm{cr}})}{1-\gamma_{\mathrm{cr}}} and χ(γ)\chi(\gamma) is the BFKL characteristic function. The critical line (z=0z = 0) is shown in red. The initial condition for scattering is given at ξs=0\xi_s = 0 and for heavy nuclei the initial condition is shown by the blue line. At the vertical blue dotted line ξ=ξ0A\xi= \xi_0^A the amplitudes described in region I and II are matched.

The kinematic region where we are looking for the solution for the scattering amplitude of a dipole (x10=rx_{10}=r) with a nucleus target in the plot with ξs\xi_s and ξ\xi axes is shown in Figure 1, where ξ=ln⁡(x102Qs2(Y=YA,b))\xi=\ln(x_{10}^{2}Q_s^{2}(Y=Y_A,b)) and ξs=ln⁡(Qs2(Y)/Qs2(Y=YA,b))\xi_s=\ln(Q_s^{2}(Y)/Q_s^{2}(Y=Y_A,b)). One can see that for z<0z<0 we have the perturbative QCD region where the non-linear corrections are small and we can safely use the BFKL linear equation for the scattering amplitude. The geometrical scaling variable zz is defined as z=ln⁡(r2Qs2(δY~,b))=ξs+ξz=\ln(r^{2}Q_s^{2}(\delta\widetilde{Y},b))=\xi_s+\xi, with the saturation moment Qs2(Y)=Qs2(YA)eαˉSκ(Y−YA)Q_s^{2}(Y)=Q_s^{2}(Y_A)e^{\bar{\alpha}_S\kappa(Y-Y_A)}, κ=4.88\kappa=4.88, and YAY_A denotes YA=ln⁡A1/3Y_A=\ln A^{1/3}, where AA is the number of nucleons in a nucleus.

For z>0z>0 the non-linear corrections become essential and we enter the saturation region. For the scattering with nuclei, the saturation region can be divided in two parts. For ξ<ξ0A\xi<\xi_0^A the amplitude has the geometric scaling behaviour [1] and depends only on one variable, zz. For ξ>ξ0A\xi>\xi_0^A this geometric scaling behaviour is broken. This process can be characterized by two kinematic regions: for r⊥Qs(Y0)<1r_\perp Q_s(Y_0)<1 and r⊥Qs(Y−Y0)<1r_\perp Q_s(Y-Y_0)<1 we can replace NN by the BFKL Pomeron. For r⊥Qs(Y0)>1r_\perp Q_s(Y_0)>1 and r⊥Qs(Y−Y0)<1r_\perp Q_s(Y-Y_0)<1 the elastic amplitude is in the saturation region and the production of gluons can be computed using the BFKL Pomeron exchange. Finally, r⊥Qs(Y0)>1r_\perp Q_s(Y_0)>1 and r⊥Qs(Y−Y0)>1r_\perp Q_s(Y-Y_0)>1 is the kinematic region where non-linear corrections for gluon production are essential.

It turns out that Eq. (1) can be rewritten in a simple form introducing a new function,

N(z,δY~,δY0)=2N(z,δY~)−ND(z,δY~,δY0),(3)\mathcal{N}(z,\delta\widetilde{Y},\delta Y_0)=2\mathcal{N}(z,\delta\widetilde{Y})-\mathcal{N}^{D}(z,\delta\widetilde{Y},\delta Y_0), \tag*{(3)}

where N(z,δY~)\mathcal{N}(z,\delta\widetilde{Y}) is the solution of the BK equation. The new variables δY~\delta\widetilde{Y} and δY0\delta Y_0 are defined as δY~=αˉS(Y−YA)\delta\widetilde{Y}=\bar{\alpha}_S(Y-Y_A) and δY0=αˉS(Y0−YA)\delta Y_0=\bar{\alpha}_S(Y_0-Y_A). This function has clear physics meaning: the inelastic cross section of all events with rapidity gap from Y=0Y=0 to Y=Y0Y=Y_0. Then, Eq. (1) in terms of the function N\mathcal{N} takes the form of the BK equation, viz.

∂N01∂Y=αˉS∫d2x022πx012x022x122{N02+N12−N02N12−N01},(4)\frac{\partial\mathcal{N}_{01}}{\partial Y}=\bar{\alpha}_S\int\frac{d^2x_{02}}{2\pi}\frac{x_{01}^{2}}{x_{02}^{2}x_{12}^{2}}\left\{\mathcal{N}_{02}+\mathcal{N}_{12}-\mathcal{N}_{02}\mathcal{N}_{12}-\mathcal{N}_{01}\right\}, \tag*{(4)}

with the initial conditions N(z→z0,δY~=δY0,δY0)=2N(z0,δY0)−N2(z0,δY0)\mathcal{N}(z\to z_0,\delta\widetilde{Y}=\delta Y_0,\delta Y_0)=2\mathcal{N}(z_0,\delta Y_0)-\mathcal{N}^{2}(z_0,\delta Y_0); it holds only in the region I, while in region II we have to use a more general expressions for the elastic scattering amplitudes (see Ref. [1]).

Our strategy for finding solution looks as follows: first we solve Eq. (4) with N01=1−ΔD\mathcal{N}_{01}=1-\Delta^{D}, and after that we return to Eq. (1).

Modified homotopy approach for Δ0D\Delta^{D}_{0} and numerical estimation for Δ1D\Delta^{D}_{1}

Including part of the non-linear term into the definition of L[u0]\mathcal{L}[u_{0}], we can find the solution to non-linear equation of Eq. (4), suggested in Refs. [1]. Following the main ideas of Ref. [6], we solve Eq. (4) replacing N(z,δY0)\mathcal{N}(z,\delta Y_{0}) by 1−Δ0D(z,δY0)1-\Delta^{D}_{0}(z,\delta Y_{0}). For this function the equation takes the form

∂Δ01D∂Y=αˉS∫d2x022πx012x022x122{Δ02DΔ12D−Δ01D}(5)\frac{\partial\Delta^{D}_{01}}{\partial Y}=\bar{\alpha}_{S}\int\frac{\mathrm{d}^{2}x_{02}}{2\pi}\frac{x_{01}^{2}}{x_{02}^{2}x_{12}^{2}}\left\{\Delta^{D}_{02}\Delta^{D}_{12}-\Delta^{D}_{01}\right\} \tag*{(5)}

with the corresponding initial conditions for Δ01D\Delta^{D}_{01}. We suggest to simplify the non-linear term replacing it by

∫d2x022πx012x022x122Δ02DΔ12D→Δ01D∫0zdz′Δ02D=Δ01D(ζ−∫z∞dz′Δ02D).(6)\int\frac{\mathrm{d}^{2}x_{02}}{2\pi}\frac{x_{01}^{2}}{x_{02}^{2}x_{12}^{2}}\Delta^{D}_{02}\Delta^{D}_{12}\to\Delta^{D}_{01}\int_{0}^{z}\mathrm{d}z'\Delta^{D}_{02}=\Delta^{D}_{01}\left(\zeta-\int_{z}^{\infty}\mathrm{d}z'\Delta^{D}_{02}\right). \tag*{(6)}

This contribution stems from the region x02≪x01x_{02}\ll x_{01} (see Ref. [6]) and, to find the solution Δ0D\Delta^{D}_{0}, we need to solve the equation L(Δ0D)=0\mathcal{L}(\Delta^{D}_{0})=0. Introducing Δ0D(z,ξs)=exp⁡(−Ω(0)(z,ξs))\Delta^{D}_{0}(z,\xi_{s})=\exp\left(-\Omega^{(0)}(z,\xi_{s})\right), we can obtain the next general equation (ξs=κ δY~)(\xi_{s}=\kappa\,\delta\widetilde{Y}),

κ∂2Ω(0)∂ξs ∂z=1−e−Ω(0)(z,ξs),(7)\kappa\frac{\partial^{2}\Omega^{(0)}}{\partial\xi_{s}\,\partial z}=1-e^{-\Omega^{(0)}(z,\xi_{s})}, \tag*{(7)}

which in the region I has traveling wave solution (Ref. [7]) with the geometric scaling behaviour, that can be found from the following implicit equation:

U(Ω(0,I),Ω0(0)=a)=∫Ω0(0)Ω(0)dΩ′Ω′+exp⁡(−Ω′)−Ω0(0)=2κ(z+C2),(8)\mathcal{U}\left(\Omega^{(0,I)},\Omega^{(0)}_{0}=a\right)=\int_{\Omega^{(0)}_{0}}^{\Omega^{(0)}}\frac{\mathrm{d}\Omega'}{\sqrt{\Omega'+\exp(-\Omega')-\Omega^{(0)}_{0}}}=\sqrt{\frac{2}{\kappa}}(z+C_{2}), \tag*{(8)}

C2C_{2} as well as Ω0(0)\Omega^{(0)}_{0} can be found using the boundary conditions. We can solve the Eq. (8) for Ω(0,I)(z)\Omega^{(0,I)}(z), and obtain Δ0(0,I)\Delta^{(0,I)}_{0} which is given by

Δ0(0,I)(z)=ΔLT(z)exp⁡(−a+∑k=1∞(−1)k−1(2k−1)!!(2κ)k+122kk k!(z−z~)2(k+12)ΔLTk(z)).(9)\Delta^{(0,I)}_{0}(z)=\Delta_{LT}(z)\exp\left(-a+\sum_{k=1}^{\infty}\frac{(-1)^{k-1}(2k-1)!!(2\kappa)^{k+\frac{1}{2}}}{2^{k}k\,k!(z-\widetilde{z})^{2(k+\frac{1}{2})}}\Delta^{k}_{LT}(z)\right). \tag*{(9)}

where ΔLT(z)=exp⁡(−(z−z~)22κ)\Delta_{LT}(z)=\exp\left(-\frac{(z-\tilde{z})^2}{2\kappa}\right) is the Levin-Tuchin solution [6]. In the region II, Ω(0,II)(z,ξs)\Omega^{(0,II)}(z,\xi_s) is described by the equation

∂2Ω(0,II)(z,t)∂z2−∂2Ω(0,II)(z,t)∂t2=1κ(1−e−Ω(0,II)(z,t))(10)\frac{\partial^2\Omega^{(0,II)}(z,t)}{\partial z^2}-\frac{\partial^2\Omega^{(0,II)}(z,t)}{\partial t^2}=\frac{1}{\kappa}\left(1-e^{-\Omega^{(0,II)}(z,t)}\right) \tag*{(10)}

where z=ξs+ξz=\xi_s+\xi and t=ξs−ξt=\xi_s-\xi, whose solution is given again by the implicit traveling wave solution U(Ω(0,II),Ω0)U\left(\Omega^{(0,II)},\Omega_0\right) (see Eq. (8)). Therefore for Δ(0,II)(z,t)\Delta^{(0,II)}(z,t) we have

Δ(0,II)=ΔLTexp⁡(−a+∑k=1∞(−1)k−1(2k−1)!!(2κ)k+122kk!((1+ν)z+νt−z^)2(k+12)ΔLTk)(11)\Delta^{(0,II)}=\Delta_{LT}\exp\left(-a+\sum_{k=1}^{\infty}\frac{(-1)^{k-1}(2k-1)!!(2\kappa)^{k+\frac{1}{2}}}{2^k k!((1+\nu)z+\nu t-\hat{z})^{2(k+\frac{1}{2})}}\Delta_{LT}^{k}\right) \tag*{(11)}

Now, we are in condition to use the homotopic approach to obtain the first correction Δ1D\Delta_1^D from the general equation valid in the region I and II:

(κ∂∂z+z)Δ1D(z,ξ,z0)=−NL[Δ0D](12)\left(\kappa\frac{\partial}{\partial z}+z\right)\Delta_1^D(z,\xi,z_0)=-\mathcal{N}_{\mathcal{L}}[\Delta_0^D] \tag*{(12)}

where the non-linear contribution is given by

NL[ΔD]=αˉS∫d2x022πx012x022x122Δ0D(x02)Δ0D(x12)−Δ0D∫x012dx022x022Δ2D(13)\mathcal{N}_{\mathcal{L}}[\Delta^D]=\bar{\alpha}_S\int\frac{d^2x_{02}}{2\pi}\frac{x_{01}^2}{x_{02}^2x_{12}^2}\Delta_0^D(x_{02})\Delta_0^D(x_{12})-\Delta_0^D\int^{x_{01}^2}\frac{dx_{02}^2}{x_{02}^2}\Delta_2^D \tag*{(13)}

For numerical estimation we have to insert Δ0D\Delta_0^D in the non-linear term and the particular solution in the region I with geometric scaling Δ1D\Delta_1^D is given by

Δ1D(z,z0)=−Δ0D(z,z0)∫z∞dz′1Δ0D(z′,z0)NL[Δ0D(z′)](14)\Delta_1^D(z,z_0)=-\Delta_0^D(z,z_0)\int_z^\infty dz'\frac{1}{\Delta_0^D(z',z_0)}\mathcal{N}_{\mathcal{L}}[\Delta_0^D(z')] \tag*{(14)}

In Figure 2 we present the numerical estimates for NL\mathcal{N}_{\mathcal{L}} and the ratio Δ1D/Δ0D\Delta_1^D/\Delta_0^D turns out to be small. In the region II Δ1D(z,ξ,z0)\Delta_1^D(z,\xi,z_0) has been found in our previous paper [1] (see Eq. 58-67) and NL[ΔD]\mathcal{N}_{\mathcal{L}}[\Delta^D] has the same form as Eq. (13) with Δ0(0,II)\Delta_0^{(0,II)}. Finally, the solution Δ1D(z,ξ,z0)\Delta_1^D(z,\xi,z_0) has the following form:

Figure 2. Fig. 2-a describes the ratio Nc[Δ0D]/Δ0D\mathcal{N}_c[\Delta_0^D]/\Delta_0^D. Fig. 2-b describes Δ1D/Δ0D\Delta_1^D/\Delta_0^D versus zz for z>z0z > z_0. z0z_0 is taken to be 3, αˉS=0.2\bar{\alpha}_S = 0.2.

Δ1D(z,ξ,z0)∼Δ0D(z,ξ,z0)(−erf⁡(z0−4ln⁡22κ)+erf⁡(z−4ln⁡22κ))(15)\Delta_1^D(z,\xi,z_0)\sim\Delta_0^D(z,\xi,z_0)\left(-\operatorname{erf}\left(\frac{z_0-4\ln2}{\sqrt{2\kappa}}\right)+\operatorname{erf}\left(\frac{z-4\ln2}{\sqrt{2\kappa}}\right)\right) \tag*{(15)}

which vanishes at z→z0z\to z_0 and in the region of large zz this contribution is rather small.

Conclusions

Using the modified homotopy approach we found that the first iteration of the homotopy approach gives the main contribution in both kinematic regions which we consider for the diffractive production. We also found that for ξ<ξ0A\xi< \xi_0^A our solution shows the geometric scaling behaviour, while for ξ>ξ0A\xi> \xi_0^A this behaviour is strongly violated. We found that the analytical solution of the non-linear equation, reproduces the initial and boundary conditions. The second iteration with zero initial and boundary conditions turns out to be small and could be taken into account together with higher iterations using the regular perturbative procedure.

Acknowledgments

This research was supported by Fondecyt (Chile) grant No. 1231829. J.G. express his gratitude to the PhD scholarship USM-DP No. 029/2024.

References

  1. [1]C. Contreras, E. Levin, and R. Meneses, Phys. Rev. D, 107(9):094030, 2023.
  2. [2]C. Contreras, J. Garrido, E. Levin, and R. Meneses, Phys. Rev. D, 110(5):054045, 2024.
  3. [3]I. Balitsky, Nucl. Phys. B, 463:99, 160, 1996.
  4. [4]Y.V. Kovchegov, Phys. Rev. D, 60:014020, 034008, 1999.
  5. [5]Y. V. Kovchegov and E. Levin, Nucl.Phys. B, 577:221, 239, 2000.
  6. [6]E. Levin and K. Tuchin, Nucl.Phys. B, 573:833, 852, 2000.
  7. [7]A.D. Polyanin and V.F. Zaitsev, CRC Press, 2004.

Paper details

vlnc:W_jjx2cfa9
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents