Comment on "Quantitative Condition is Necessary in Guaranteeing the Validity of the Adiabatic Approximation" [arXiv:1004.3100]
aa r X i v : . [ qu a n t - ph ] A p r Comment on ”Quantitative Condition is Necessary inGuaranteeing the Validity of the Adiabatic Approxima-tion”
Recently, the authors of Ref.[1] claimed that they haveproven the traditional adiabatic condition is a necessary con-dition. Here, it is claimed that there are some mistakes and anartificial over-strong constraint in [1], making its result incon-vincible.In their proof in [1], the author underestimated the contri-butions of other small components in a general adiabatic evo-lution, which is implied in Eq.(7) in [1]. A detailed version ofEq.(7) in [1] is as follows | ψ ( t ) i = β ( t ) | ψ adi ( t ) i + δ ( t ) | ψ adi ⊥ ( t ) i , (1)where | δ ( t ) | ≪ β ( t ) + | δ ( t ) | =
1. In the following,time parameter t is omitted for convenience. Actually, Eq.(1)is equivalent to Eq.(10.54) in [2]. Differentiating both sides ofEq.(1) we get | ˙ ψ i = β | ˙ ψ adi i + ˙ β | ψ adi i + ˙ δ | ψ adi ⊥ i + δ | ˙ ψ adi ⊥ i . (2)Substituting Eq.(1) into the Schrodinger’s equation, and com-bining the result with Eq.(2), we get( ~ = i | ˙ ψ adi i = H | ψ adi i − i ˙ ββ | ψ adi i + β (cid:0) H δ | ψ adi ⊥ i− i ˙ δ | ψ adi ⊥ i − i δ | ˙ ψ adi ⊥ i (cid:1) . (3)As ˙ β = − | δ | β d | δ | dt , a term which contains ˙ β is subleading in theright side of Eq.(3). To simplify the following discussion wemay neglect its effect. So Eq.(6) in [1] declared to be a neces-sary condition is valid if and only if (cid:13)(cid:13)(cid:13)(cid:13) β ( − i ˙ δ | ψ adi ⊥ i − i δ | ˙ ψ adi ⊥ i ) (cid:13)(cid:13)(cid:13)(cid:13) is small comparing to (cid:13)(cid:13)(cid:13) H | ψ adi i (cid:13)(cid:13)(cid:13) . Here and thereafter the sym-bol k · k denotes the norm of vector. This generally requires (cid:12)(cid:12)(cid:12) ˙ δ (cid:12)(cid:12)(cid:12) ≪ E m and (cid:13)(cid:13)(cid:13) δ | ˙ ψ adi ⊥ i (cid:13)(cid:13)(cid:13) ≪ E m which are extra requirementsin addition to the existing requirement of Eq.(7) in [1].It’s explicit that the ’Proof’ in [1] can not be completedwithout the extra limitation Eq.(6) in [1]. Substituting Eq.(1)and Eq.(2) into Eq.(12) in [1] and neglect the term containing˙ β , we get c m = E m − E n [ i βγ + ( i ˙ δ − E n δ ) γ + i δγ )] , (4)where γ = h E m | ˙ ψ adi i , γ = h E m | ψ adi ⊥ i , γ = h E m | ˙ ψ adi ⊥ i . Termscontaining δ can be ignored assuming δ is small enough, c m = E m − E n ( i βγ + i ˙ δγ ) , (5)but terms containing ˙ δ cannot be ignored [2]. ComparingEq.(5) here with Eq.(13) in [1], one can see that the proof in [1] can only be applied to special systems. It is quite possi-ble that neither of the two terms in the right side of Eq.(5) issmall but their summation is small. In this case, quantitativecondition is unnecessary. Specifically, the second example insection V of [3] supports the conclusion.Moreover, there is a loophole in the logic of the ”proof” in[1]. E n in its Eq.(11) can be replaced by any number otherthan E m to complete its following ”proof”. If a replacementis done, then nothing useful can be deduced by its ”proof”.The loophole arises from the suspicious ” ≃ ” in Eq.(13) in[1]. In fact, the second term in the right side of Eq.(12) in [1], h E m | E n | ψ i E m − E n = E n c m E m − E n , is at the same order of (or even much largerthan) c m . It can not be simply substituted by h E m | E n | ψ adi i E m − E n = † and Jianda Wu ⋆ Department of Modern Physics, University of Science andTechnology of China, Hefei 230026, People’s Republic ofChina; Department of Physics & Astronomy, Rice Univer-sity, Houston, Texas 77005, USAPACS number: 03.65.Ca, 03.65.Ta, 03.65.Vf. † [email protected] ⋆ [email protected]@rice.edu