Theory of spin precession
monitored by laser
pulse
Abstract
We first predict the splitting of a spin degenerate impurity level when this impurity is irradiated by a circularly polarized laser beam tuned in the transparency region of a semiconductor. This splitting, which comes from different exchange processes between the impurity electron and the virtual pairs coupled to the pump beam, induces a spin precession around the laser beam axis, which lasts as long as the pump pulse. It can thus be used for ultrafast spin manipulation. This effect, which has similarities with the exciton optical Stark effect we studied long ago, is here derived using the concepts we developed very recently to treat many-body interactions between composite excitons and which make the physics of this type of effects quite transparent. They, in particular, allow to easily extend this work to other experimental situations in which a spin rotates under laser irradiation.
PACS.: 71.35.-y Excitons and related phenomena
Long ago, Danièle Hulin and her group [1] discovered that, when a semiconductor is irradiated by photons with energy too low to create electron-hole pairs, the exciton line blue-shifts. We have shown [2] that this shift, which disappears when the pump laser is turned off, comes from interactions between the real exciton created by the probe photon and the virtual excitons coupled to the pump beam [3].
In this communication, we predict an effect which has similarities with this exciton optical Stark shift: When an impurity is irradiated by a pump beam tuned in the transparency region of a semiconductor, its electronic levels shift: The electron bound to the donor interacts with the virtual electron-hole pairs coupled to the pump beam, either by Coulomb interaction, or by Pauli exclusion. If we choose the pump polarization in such a way that the exchange processes between the virtual pair and the up and down electrons of the impurity are different, this Pauli “interaction” splits the impurity level. As a result, the spin of the impurity electron precesses around the laser beam axis, as long as the pump is turned on. This effect can thus be used for ultrafast spin manipulation, a subject of great technological interest in the present days [4-10].
The impurity shift induced by a pump beam is derived following a procedure inspired from the one we used long ago to get the exciton optical Stark shift [3]. However, to enlighten the physics of this effect, we here calculate it using a “commutation technique” similar to the one we recently developed for excitons interacting with excitons [11] and which allows to identify the two basic ingredients of the electron-virtual pair interactions, namely a direct Coulomb scattering and a Pauli (or exchange) “scattering” — without any Coulomb contribution. The shift results from the interplay between the two, while the splitting only comes from different carrier exchanges.
To make the physics which controls the impurity level shift more transparent, we, in the first part, assume that the impurity electron and the electron of the virtual pairs have the same spin. The spin degrees of freedom and the laser polarization, of course crucial to get an impurity level splitting, will be introduced in the second part.
We end this communication by reconsidering other experimental conditions in which spins can rotate under laser pulse, namely free electron in a quantum well [8,9] and electrons trapped in quantum dots [5]. We explicitly show how our present theory can be easily extended to these cases.
Impurity level shift under laser irradiation
Let us consider a semiconductor having a ionized donor. Its Hamiltonian reads , where is the bare semiconductor Hamiltonian, with and , while is the Coulomb interaction between the ionized donor and the carriers. This interaction, , reads in second quantization,
| (1) |
and are the creation operators for free electrons and holes, i. e., , while or are the Coulomb matrix elements between free carriers, in 3D or 2D systems. In the presence of the ionized donor, the one-electron eigenstates, , can be formally written as
| (2) |
If we now irradiate this system with pump photons , the coupled matter-photon Hamiltonian reads , where is the bare photon Hamiltonian and the semiconductor-photon coupling. , which creates one electron-hole pair with momentum , can be written as where , with , is the creation operator for one free electron-hole pair [12] with center of mass momentum and relative motion momentum . It is such that , where is the band gap, while , with and , is the pair energy.
As for the exciton optical Stark effect [3], the impurity level shift results from the difference between the impurity level change and the vacuum level change induced by the pump beam. For , the eigenstate with a ionized impurity and photons is , its energy being . At lowest order in , this energy becomes
| (3) |
In a similar way, for , the eigenstates with one electron and photons are , their energy being , while at lowest order in , they read
| (4) |
The impurity level shift induced by the pump beam, , is thus
| (5) |
This quantity, linear in the pump intensity, is formally similar to our expression of the exciton optical Stark shift, with and just replacing the probe exciton creation operator and energy .
In order to calculate , we introduce the Coulomb creation potential , which, in this problem, is defined as . It precisely reads
| (6) |
From the formal definition of , it is easy to check that
| (7) |
which is valid for any scalar . This allows to split as
| (8) |
Note that we have done a similar splitting in the case of the exciton optical Stark shift [3].
As explicitly shown below, , given by
| (9) |
in which we have set , comes from “Pauli interaction” between the impurity electron and the virtual pair. On the opposite, and , given by
| (10) |
contain one or two operators, so that they come from Coulomb interaction between the impurity electron and the virtual pair.
In the following, it will be convenient to develop on free pair states, according to
| (11) |
| (12) |
“Commutation technique” for a free pair interacting with an impurity electron
and are easy to write in terms of the two “scatterings” controling the physics of this problem, namely and , which appear in a “commutation technique” inspired from the one we recently developed to treat many-body effects between composite excitons [11]. From
| (13) |
one of these two “scatterings”, which is dimensionless, is found to be
| (14) | |||||
It corresponds to a bare electron exchange between the impurity level and the pair , which transforms them into an “out” impurity level and an “out” pair . Note that this scattering is Coulomb free. From eq. (13), we can show that
| (15) |
The second scattering, defined through
| (16) |
is found to be
| (17) |
where . It corresponds to direct Coulomb interactions between the impurity electron and the pair, without any carrier exchange.
Calculation of the impurity level shift
(i) Pure Pauli term
Equations (9,11,15) allow to write as
| (18) |
Equation (18) makes clear that this part of the shift is linked to “Pauli interaction”, i. e., exchange between the impurity electron and the virtual pairs. By noting that, at large detuning , tends to , we can extract this limit from to write it as , where, due to eqs. (12,14), is precisely given by
| (19) |
For large , the bracket of eq. (19) tends to zero, so that does reduce to . This limit has to be compared to the one of the similar Pauli term in the optical Stark shift, namely . The link between these two limits can be physically understood by noting that, in the case of the exciton shift, a virtual pump pair can exchange both, its electron and its hole, with the probe exciton, while here, it can only exchange its electron with the impurity level: The numerical prefactor of the large detuning leading term just results from one carrier exchange instead of two.
In order to calculate the next order term , we use
| (20) |
which follows from . The first term of eq. (20) leads to replace by in eq. (19). As , while for bound states for , we find that the contribution of this term to is of the order of , where . A similar behavior is found for each of the two terms of . On the opposite, the term of eq. (20), which corresponds to Coulomb interaction inside the virtual pair (see fig. 1b), becomes singular for large momentum transfers. It leads to
| (21) |
with for 3D and for 2D; so that we end with
| (22) |
Note that, as , is in fact the dimensionless parameter associated to a Coulomb expansion.
(ii) First order Coulomb term between the impurity electron and the virtual pairs
We now turn to . Using eqs. (10,11,15,16), it reads
| (23) |
where is the sum over of . Being made of a direct Coulomb process between the impurity electron and the pair, followed by an electron exchange (see fig. 1c), is actually an exchange Coulomb scattering.
To get the lowest order term in , i. e., in Coulomb interaction, we can replace by its free carrier expression , i. e., by . The two terms of being then equal, we are left with the exchange term, which gives
| (24) | |||||
(iii) Correlation term
The last contribution contains two Coulomb interactions between the impurity electron and the free pair, i. e., two () factors. In the large detuning limit, it thus behaves as at least (other factors possibly appearing if we expand according to eq. (7)). Consequently, in this large detuning limit, is negligible in front of and . On the opposite, the contribution is the one possibly leading to resonances in the impurity level shift. Indeed, if we look at eq. (10), we see that contains acting on two electrons plus one hole. The corresponding eigenstates being the excitons bound to an impurity, we can inject the closure relation for these states in front of this dependent operator. then shows poles at , where the ’s are the energies of these excitons bound on impurity.
This leads us to conclude that, at large detuning, the impurity level shift is entirely controlled by electron exchange between the impurity and the virtual pairs coupled to the pump beam, without any Coulomb contribution (see fig. 1a). The next order term, which is smaller, is also due to an electron exchange but contains, in addition, one Coulomb interaction, either inside the virtual pairs as in (see fig. 1b), or between these virtual pairs and the impurity electron as in (see fig. 1c). On the opposite, possible resonances at the bound exciton energies can be found in the correlation term , which, at large detuning, gives a negligible contribution.
Impurity level splitting
Let us now see how the pump polarization and the spin degrees of freedom affect these results.
The semiconductor-photon coupling now reads
| (25) |
where creates a pair with an electron spin and a hole momentum for bulk materials while only for quantum wells. The ’s depend on photon polarization. For bulk materials, their non-zero values are and , while, for quantum wells, these ’s are zero . In the case of a circularly polarized beam , the ’s are such that and , while for a linear beam along (resp. ), they are (resp. ).
In addition to these complexities in the semiconductor-photon interaction, we have also to take into account the fact that the impurity levels are now degenerate, the up and down spins having the same energy — in the absence of pump beam. Consequently, it is now necessary to use degenerate perturbation theory to get the impurity level change induced by the laser beam. It is possible to show that this change is obtained from the diagonalization of a matrix, its eigenvalues being
| (26) |
| (27) |
By taking into account the vacuum level change induced by the pump beam, still given by eq. (3), we end with an impurity level having an average shift equal to , and a splitting given by , where
| (28) |
Note that eq. (28) is a generalization of eq. (5), in the presence of spin degrees of freedom.
To get these , we use a commutation technique similar to the one without spin. In the presence of spins, the three scatterings , and are now the product of an orbital part, which is the one without spin, and a spin part. Due to spin conservation in Coulomb and exchange processes, this spin part is just for the direct scattering (see fig. (1d), and for the exchange scatterings and (see fig. 1e ).
It is then easy to show that, again, the large detuning leading term of is entirely controlled by electron exchange between the impurity and the virtual pairs coupled to the pump, the next order term having just one additional Coulomb interaction either inside the pair or between the pair and the impurity electron. The two first terms of correspond to the two first terms of as obtained previously in eqs. (18) and (23), with just replaced by
| (29) |
We can then note that , since for a given , there is only one which makes , while is equal to for bulk samples, and for quantum wells. This shows that, when the pump beam is linear, so that : The impurity level has a blue shift equal to for quantum wells, and for bulk materials, but no splitting. On the opposite, for circular beams, , so that the impurity level splits: One impurity level blue shifts of an amount , while the other is unchanged for quantum wells, or shifted by for bulk materials. The splitting is then either or .
Spin precession of an impurity electron induced by a pump beam
Let us take as initial impurity state. If we turn on a circularly polarized pump beam which propagates along , the up and down spins are shifted differently, due to their different electron exchanges with the virtual pairs, so that becomes
| (30) |
within a phase factor, being the shift between the impurity electrons calculated previously. The projections of over and staying unchanged, the spin of the impurity electron thus precesses around the axis with a period . Since is of the order of — which is just the exciton optical Stark shift, within a factor , in the large detuning limit — , we thus expect a precession period of the order of 1psec within the experimental conditions giving an exciton optical Stark shift of the order of 1mev. We can note that this period is far shorter than the spin relaxation time, which is of the order of 1nsec.
Extension of the theory to other spin precessions
Let us end this communication by considering two cases in which spin precession induced by laser beams has been described.
(i) Free electron in a quantum well [8,9]
This case can be readily deduced from the above results by setting the Coulomb potential between the carriers and the ionized impurity equal to zero. This leads to replace by , by and by , with , in the formal expression of the shift as well as in its , and contributions. We have shown that, in the large detuning limit, the two first terms of the shift are controlled by an electron exchange, with possibly one Coulomb interaction inside the virtual pairs or between these pairs and the impurity electron, the Coulomb interaction with the ionized donor entering at the next order, , only. This shows that the shift and splitting of the impurity electron and the ones of a free electron are thus just the same for these two large detuning terms, provided that , for eq. (21) to be valid. On the opposite, the possible resonances coming from the contribution differ. They are now controlled by the two electron-one hole eigenstates, i. e., the trions, while, in the presence of impurity, they are controlled by excitons bound to the impurity. As the coupling between photon and trion is in fact extremely weak in the large sample limit [13], the weights of these resonances are expected to be rather small.
(ii) Electron in a quantum dot [5]
The spin precession of an electron trapped in a quantum dot can also be deduced from the above theory. The one-body electron Hamiltonian has just to now include the dot confinement. Instead of , the creation operator for a Coulomb free electron reads , with . These eigenstates a priori include bound states as well as extended states, if the barrier height is finite.
If we now consider one electron trapped in the dot ground state , its shift is given by eq. (5), with and replaced by and , while the dot-photon coupling has now to be written as . This shift can be calculated using a “commutation technique” formally similar, the Coulomb creation potential now reading
| (31) |
where and are the Coulomb matrix elements between dot states and .
The calculation of which is performed in a quite similar way, shows that, at large detuning, the shift is again controlled by electron exchange, its leading term now reading , while resonant contributions in the correlation term must appear at the eigenenergies of a “trion” in the dot.
Conclusion
We have shown that, due to carrier exchanges between the impurity and the virtual pairs coupled to a pump beam tuned in the transparency region of a semiconductor, the up and down electronic levels of an impurity blue shift. The degenerate levels of this impurity can also split if the pump beam is circularly polarized, due to differences in these carrier exchanges. This splitting induces a spin precession around the laser beam axis, which lasts as long as the pulse. It can thus be used to manipulate spins. We have also shown how the present theory can be extended to other spin precessions induced by laser beam, such as the one of free electrons in a quantum well or the one of electrons trapped in a quantum dot.
We wish to thank J. Tribollet for inducing this work and C. Mora for stimulating discussions.
REFERENCES
[1] A. Mysyrowicz, D. Hulin, A. Antonetti, A. Migus, W.T. Masselink, H. Morkoç, Phys. Rev. Lett. 56, 2748 (1986).
[2] For a review, see M. Combescot, Phys. Rep. 221, 168 (1992).
[3] M. Combescot, R. Combescot, Phys. Rev. Lett. 61, 117 (1988); Phys. Rev. B 40, 3191 (1989). M. Combescot, Phys. Rev. B 41, 3517 (1990).
[4] D. Loss, D.P. Di Vincenzo, Phys. Rev. A 57, 120 (1998).
[5] A. Imamoglu, D.D. Awschalom, G. Burkard, P.D. Di Vincenzo, D. Loss, M. Sherwin, A. Small, Phys. Rev. Lett. 83, 4204 (1999).
[6] P. Chen, C. Piermarocchi, L.J. Sham, Phys. Rev. Lett. 87, 67401 (2001).
[7] C. Piermarocchi, P. Chen, L.J. Sham, D.G. Steel, Phys. Rev. Lett. 89, 167402 (2002).
[8] J.A. Gupta, R. Knobel, M. Samarth, D.D. Awschalom, Science 292, 2459 (2002).
[9] J. Bao, A.V. Bragas, J.K. Furdyna, R. Merlin, Nat. Mater. 2, 175 (2003).
[10] P. Chen, C Piermarocchi, L.J. Sham, D. Gammon, D.G. Steel, Phys. Rev. B 69, 75320 (2004).
[11] M. Combescot, O. Betbeder-Matibet, Europhys. Lett. 58, 87 (2002). O. Betbeder-Matibet, M. Combescot, Eur. Phys. J. B 27, 505 (2002).
[12] While the semiconductor-photon interaction can also be written in terms of excitons, it is in fact more convenient to write it in terms of free electron-hole pairs for problems dealing with large detunings.
[13] M. Combescot, J. Tribollet, Solid State Com. 128, 273 (2003).