Conversion to cell from double is not possible in for loop

I am trying to write a code. Following is the part of code relative to query

Energy= En + (AV*q) ; 
Widths = [ 10*AA, 10*AA, 10*AA ] ;
Entries= 3;
xmax=0;
for n=1: Entries
xmax = xmax + Widths ( n ) ;
end
Slope = AV*eV / xmax ;
Pots = [ 0.7*eV , 0*eV , 0.7*eV ] ;
Pot0Left = AV*q ; 
Pot0Right= 0*eV ; 
Precision = Prec ; 
kr = sqrt(2*me*(Pots(1) + Pot0Left - Energy))/hbar ;
kl = sqrt (2*me*(Energy-Pot0Left))/hbar ;
SysMat(1,1)=1- 1i*kr/kl ; SysMat (1,2)=1+ 1i*kr/kl ;
SysMat(2,1)=1+ 1i*kr/kl ; SysMat (2,2)=1- 1i*kr/kl ;
SysMat=0.5*SysMat ;
for n=1: Entries
xposl = 0 ;
for j =1 :(n-1)
xposl = xposl + Widths (j) ;
end
DivW = Widths(n)/Precision ; 
potl = Pots (n) + AV*q - xposl * Slope ;
for l=1:Precision
potr = potl - DivW * Slope ;
if (mod(n,2)==1)
a = sqrt(2*me*(potl - Energy))/hbar ;
M(1,1) = exp(-a*DivW); 
M(1,2) = 0 ;
M(2,1) = 0 ; 
M(2,2) = exp(a*DivW) ;
else 
k = sqrt (2*me*(Energy - potl))/hbar ;
M(1,1) = exp( -1i*k*DivW ) ; M( 1 , 2 ) = 0 ;
M(2,1) = 0 ; M(2,2) = exp( 1i*k*DivW ) ;
end
SysMat = M*SysMat ;
if (l<Precision )
if (mod(n,2)==1) 
k1= sqrt(2*me*(potr - Energy))/hbar ; 
k2= sqrt(2*me*(potl - Energy))/hbar ; 
else 
k1= sqrt (2*me*(Energy - potr))/hbar ; 
k2= sqrt (2*me*(Energy - potl))/hbar ; 
end
M(1,1)=k2+k1 ; M(1,2) = k2-k1 ;
M(2,1)=k2-k1 ; M(2,2) = k2+k1 ;
SysMat = (0.5*M*SysMat)/k2 ;
potl = potr ;
end
end 

But at the line

M(1,1) = exp(-a*DivW);

It shows the error

Conversion to cell from double is not possible.

Error in rtTransm (line 56)
M(1,1) = exp(-a*DivW);

In the workspace it shows that M is a 2X2 cell while I need it to be 2X2 double. While the other matrix SysMat is a 2X2 double.

I am varying the variable En from 0 to 10 in steps of 0.01 which in turn is varying the variable a which I think is causing the error as DivW is constant.

Any help in correcting the code is appreciated.

[Stephen23]

Preallocate M as a double:

https://www.mathworks.com/help/matlab/matlab_prog/preallocating-arrays.html

[AYUSH KUMAR]

I did that using

M = zeros([2 2])

But it is showing the error

Assignment has more non-singleton rhs dimensions than non-singleton subscripts

Error in rtTransm (line 57)
M(1,1) = exp(-a*DivW);

How can I correct this error?

Thank you for responding.

[Walter Roberson]

What is size(a) ? What is size(Divw) ?

[AYUSH KUMAR]

Since I am varying En = 0:0.01:10;

and a depends on En it is an array 0f 1X1001 double

while DivW is a constant.

[Walter Roberson]

Perhaps you should be referring to a(n) instead of to the entire vector a ?