one temperature: gamma_imp and alpha are 100.00% correlated: the data hardly determine them separately, and their errors are unreliable. Try several starting points, or fix one
joint: gamma_imp = 1.999e+12 ± 2.7e+09 s⁻¹ (truth 2e+12)
joint: alpha = 3.999e+08 ± 7e+05 s⁻¹ K⁻² (truth 4e+08)

  T   error of Γ₁ (10¹⁰ s⁻¹)
      alone  joint
 10    0.63   0.49
 30    0.74   0.46
 50    0.77   0.44
 70    0.90   0.58
 90    1.27   0.96
