External module functions dmq1 - used by mcdisp

Similarly ''going beyond'' dipolar approximation in the program mcdisp can be done with module functions dmq1 and estates. The input of dmq1 has similar arguments as du1calc, but as additional argument an orientation of the scattering vector, output should be a corresponding vector ${\mathbf m^s_{\alpha1}}(\mathbf Q)=\sqrt{(p_--p_+)}
\langle -\vert\hat \mathbf...
...Q)-\langle \hat \mathbf M_{\alpha}^{\dag }(\mathbf Q)\rangle\vert+\rangle/\mu_B$. Here $\hat \mathcal Q_{\alpha}=\frac{\hat \mathbf M_{\alpha}(\mathbf Q)}{-2 \mu_B}$ are the cartesian components of the scattering operator. dmq1 is called many times, for every scattering vector. In order to do an efficient calculation the eigenstates should be calculated only once, this is the task of function estates (see above).

The format to be used is:

extern "C" int dmq1(int & tn,double & th,double & ph,double J0,
double & J2, double & J4, double & J6,ComplexMatrix & est,double & T,
ComplexVector & mq1, float & maxE)

The meaning of the symbols is as follows:

on input
   |tn|            transition-number  
   sign(tn)        >0 standard with printouts for user information, 
                   <0 routine should omit any printout
   th              polar angle theta of the scattering vector Q 
                   (angle with the axb axis=c axis) in rad
   ph              polar angle phi of the scattering vector Q 
                   (angle with bx(axb)=a in the projection into
                   the  bx(axb),b plane = angle with a in the projection into 
				   the ab plane) in rad
   J0,J2,J4,J6     form factor functions <jn(Q)>   
   est             eigenstate matrix (as calculated by estates),
                   it should also contain population numbers of the states (row 0)
   T               Temperature[K]
   mq1(1)           ninit + i pinit (from mcdisp options  -ninit and -pinit)
   maxE            maximum transition energy (from mcdisp option maxE)
on output
   int             total number of transitions
   mq1             vector mq(alpha)=<-|-2Qalpha|+>sqrt(p- - p+)
     Note on Qalpha
        Cartesian components of the scattering operator Qalpha, alpha=1,2,3=a,b,c
        according to Lovesey Neutron Scattering equation 6.87b 
        scattering operator is given in  spherical coordinates Q-1,Q0,Q+1 (introduced
        as described above on input of th and ph) these are related to euclidean 
		components by 11.123
        the orbital and spin contributions can be given as separate 
		components  according to Lovesey Neutron Scattering 
		equations 11.55 and 11.71 (the spin part 11.71 has to be
        divided by 2), i.e.
          =<-|sum_i exp(i k ri) s_(a,b,c)|+> /2                   as defined by 11.71 / 2
          =<-|sum_i exp(i k ri) (-(k x grad_i)_(a,b,c)/|k|)|+>     as defined by 11.54 /(-|k|)
	     thus for k=0 <QS>=<S>/2 and <QL>=<L>/2 
        Q=2QS+QL, M(Q)=Q/(-2muB)=mq1/muB

The module function must perform the following tasks:

  1. for the transition number tn the vector mq1 is to be filled with the 3-components of ${\mathbf m^s_{\alpha1}}(\mathbf Q)=\sqrt{(p_--p_+)}
\langle -\vert\hat \mathbf...
...Q)-\langle \hat \mathbf M_{\alpha}^{\dag }(\mathbf Q)\rangle\vert+\rangle/\mu_B$. , i.e. for $\alpha=1,..3$ IMPORTANT: the numbering scheme of transitions has to be the same for du1calc and all the corresponding d...1 functions for observables !
  2. If the energy of this transition is zero, i.e. $\Delta(tn)=0$ (diffuse scattering), the ${\mathbf m^s_{\alpha1}}(\mathbf Q)$ would be zero because $(p_--p_+)$ vanishes (compare expression (195)). In this case the single ion module should calculate $(p_+/kT)$ instead of $(p_--p_+)$.
  3. if $T<0$ all quantities should be evaluated assuming that all Boltzmann probabilities $p_i$ are zero except for the state number $n=(-T)$, for which the probability $p_n=1$.

Martin Rotter 2017-01-10