CDF dE/dx Parameterization Studies
Presentations
Universal curve parameterizations, version 2.0
The following parameters were generated using TOF+dE/dx data and
delta cot(theta)+dE/dx from conversions. Energy loss corrections
were applied using the 2/25/2004 version of
CT_DedxAnalysis with an additional momentum-dependent correction
provided by Stefano and Vivek.
The universal curve is still parameterized by the function:
double uc(double *y,double *par) {
double c0 = par[0];
double c1 = par[1];
double b = par[2];
double a1 = par[3];
double a2 = par[4];
double c = par[5];
double mass = par[6]; // Particle mass hypothesis
double x = y[0]/mass; // y[0] is momentum, x is beta*gamma
double t = x/sqrt(1+x*x); // t is beta
return (c1*log(x/(x+b)) + c0)/(t*t) + a1*(t-1) + a2*(t-1)*(t-1) + c;
}
and the resolution on the quantity Z=log[dEdx(measured)/dEdx(predicted)] is
still parameterized by:
double res(double *y,double *par) {
double n0 = par[0];
double sigma = par[1];
double b = par[2];
double c = par[3];
double n = y[0]; // Number of dE/dx hits
double uc = y[1]; // Predicted dE/dx
return (sigma+c*(uc-15.0))*pow(n/n0,b);
}
There are separate sets of parameters for positive and negative tracks.
Universal curve parameterizations, version 1.0
The following parameters were generated using TOF+dE/dx data and
delta cot(theta)+dE/dx from conversions. Energy loss corrections
were applied using the 2/25/2004 version of
CT_DedxAnalysis.
The universal curve is parameterized by the function:
double uc(double *y,double *par) {
double c0 = par[0];
double c1 = par[1];
double b = par[2];
double a1 = par[3];
double a2 = par[4];
double c = par[5];
double mass = par[6]; // Particle mass hypothesis
double x = y[0]/mass; // y[0] is momentum, x is beta*gamma
double t = x/sqrt(1+x*x); // t is beta
return (c1*log(x/(x+b)) + c0)/(t*t) + a1*(t-1) + a2*(t-1)*(t-1) + c;
}
and the resolution on the quantity Z=log[dEdx(measured)/dEdx(predicted)] is
parameterized by:
double res(double *y,double *par) {
double n0 = par[0];
double sigma = par[1];
double b = par[2];
double c = par[3];
double n = y[0]; // Number of dE/dx hits
double uc = y[1]; // Predicted dE/dx
return (sigma+c*(uc-15.0))*pow(n/n0,b);
}
There are separate sets of parameters for positive and negative tracks.