REI TRANSDUCTION THEORY
ER 3 YVO 4 SPECTROSCOPY
3.4 Magnetic Field Orientation
0 10 20 30 40 50 60 70 80 90 B-field angle from c-axis
2 3 4 5 6 7 8
g dc
Z1
Y1
Y2
Figure 3.5: πππfor π1, π1andπ2as a function of the magnetic field angle from the c-axis.
Lastly, magnetic field orientation for the branching ratio of the optical transitions dipoles can be determined. We will use an effective spin Hamiltonian approach as previously shown in the literature that has matched quite well to their experiments [111β113]. We can use the effective spin Hamiltonian for the ground and excited state (Equation 3.1) to determine the states at a given magnetic field orientation.
First we can write down the states:
|Ξ¨+, π1β©=βπ½β |Ξ¦β3/2, π1β© +πΌβ|Ξ¦+3/2, π1β© (3.9)
|Ξ¨β, π1β©=βπΌβ|Ξ¦β3/2, π1β© +π½|Ξ¦+3/2, π1β© (3.10)
|Ξ¨+,π1β©=βπΏβ
π1|Ξ¦β3/2,π1β© +πΎβ
π1|Ξ¦+3/2,π1β© (3.11)
|Ξ¨β,π1β©=βπΏπ
1|Ξ¦β3/2,π1β© +πΎπ
1|Ξ¦+3/2,π1β© (3.12)
|Ξ¨+,π2β©=βπΏβ
π2|Ξ¦β1/2,π2β© +πΎβ
π2|Ξ¦+1/2,π2β© (3.13)
|Ξ¨β,π2β©=βπΏπ
2|Ξ¦β1/2,π2β© +πΎπ
2|Ξ¦+1/2,π2β© (3.14) where |Ξ¦π, πβ©are the labels of the states with a magnetic field along the c-axis and π is the crystal field quantum number. |Ξ¨π, πβ©is the label of the state when a static magnetic field is applied in the a-c plane. πΌ, π½, πΏπ
1, πΎπ
1, πΏπ
2, andπΎπ
2 are the state overlap coefficients.
We can then apply the transition selection rules (see Table 3.1) and transition oper- ator time-reversal symmetry properties [66, 112] to estimate the relative transition
strengths for the electric and magnetic dipole transitions forπ1-π1: π π
1, π΅β₯π,β₯ = | β¨Ξ¨+, π1|πΛπ΅β₯π|Ξ¨β,π1β© |2=|πΌπΏ+πΎ π½|2 (3.15) π π
1, π΅β₯π,β₯ = | β¨Ξ¨+, π1|πΛπ΅β₯π|Ξ¨+,π1β© |2= |πΌπΏββπ½πΎβ|2 (3.16) π π
1,πΈβ₯π,β₯ =| β¨Ξ¨+, π1|πΛπΈβ₯π|Ξ¨β,π1β© |2=|πΌπΏ+π½πΎ|2 (3.17) π π
1,πΈβ₯π,β₯ =| β¨Ξ¨+, π1|πΛπΈβ₯π|Ξ¨+,π1β© |2= |πΌπΎββ π½πΏβ|2 (3.18) π π
1, π΅β₯π,β₯ =| β¨Ξ¨+, π1|πΛπ΅β₯π|Ξ¨β,π1β© |2 =|πΌπΏβπ½πΎ|2 (3.19) π π
1, π΅β₯π,β₯ = | β¨Ξ¨+, π1|πΛπ΅β₯π|Ξ¨+,π1β© |2=|πΌπΎβ+π½πΏβ|2 (3.20) whereπΈ/π΅correspond to magnetic and electric dipoles, β₯ π/πrefers to the polar- ization of the dipole, β₯ /β₯ correspond to transitions between sign-preserving (i.e.
|Ξ¨+, π1β© β |Ξ¨+,π1β©) or sign-flipping transitions (i.e. |Ξ¨+, π1β© β |Ξ¨β,π1β©), and Λπis the transition dipole operator.
We can do the same thing forπ1-π2: π π
2,πΈβ₯π,β₯ =| β¨Ξ¨+, π1|πΛπΈβ₯π|Ξ¨β,π2β© |2 =|π½πΏ+πΌπΎ|2 (3.21) π π
2,πΈβ₯π,β₯ = | β¨Ξ¨+, π1|πΛπΈβ₯π|Ξ¨+,π2β© |2=|πΌπΏββπ½πΎβ|2 (3.22) π π
2,πΈβ₯π,β₯ =| β¨Ξ¨+, π1|πΛπΈβ₯π|Ξ¨β,π2β© |2= |πΌβπΏββπ½βπΎβ|2 (3.23) π π
2,πΈβ₯π,β₯ = | β¨Ξ¨+, π1|πΛπΈβ₯π|Ξ¨+,π2β© |2 =|πΌβπΎ+π½βπΏ|2 (3.24) π π
2, π΅β₯π,β₯ = | β¨Ξ¨+, π1|πΛπ΅β₯π|Ξ¨β,π2β© |2=|πΌβπΏβ+π½βπΎβ|2 (3.25) π π
2, π΅β₯π,β₯ =| β¨Ξ¨+, π1|πΛπ΅β₯π|Ξ¨+,π2β© |2=|πΌβπΎβ π½βπΏ|2. (3.26) We can then define the branching ratio for each dipole transition as:
π΅π, π =
π π, π ,β₯ π π, π ,β₯+π π, π ,β₯
(3.27) for transition π (i.e. π1-π1 or π1-π2) and π dipole type and polarization. The branching ratio is defined in such a way that an ideal (or even) branching ratio gives a value of 1/2 which results in even mixing between the states, while a branching ratio of 0 or 1 corresponds to no mixing.
The branching ratio for the three different possible transition dipoles for π1-π1are shown in Figure 3.6a. When the DC magnetic field is along either the c-axis or the a-axis, the branching ratio for all transitions dipoles is either 0 or 1. Only theπΈ β₯ π (ππΈ π·) transition dipole can reach an even branching ratio at a magnetic field angle of 35Β°from the c-axis. π΅||π (ππ π·) and π΅||π(ππ π·) can reach non-zero branching ratios, but the mixing is far from even (97% and 3%, respectively).
Figure 3.6: Branching ratio for a) π1-π1 and b) π1-π2 optical transitions for the different dipole operators (i.e. electric or magnetic) and orientations (i.e. parallel to the a-axis or c-axis).
Similarly, we can look at the branching ratios for π1-π2 in Figure 3.6b. Both orientation of the electric dipole can achieve an even branching ratio at a magnetic field angle 24Β° from the c-axis. The branching ratio for π΅ β₯ π (ππ π·) reaches a maximum value of 3Β·10β3which is very small.
For transduction, we care about the branching ratio, but we also need to consider the overall strength of the transition dipole. In other words, if the branching ratio is not even, but the total strength is much larger, that can still be the better transition dipole to use. We can calculate the transition dipole moments as:
π2
π, π ,β₯ =π2
π, π
π π, π ,β₯
π π, π ,β₯ +π π, π ,β₯
(3.28) π2
π, π ,β₯ =π2
π, π
π π, π ,β₯
π π, π ,β₯ +π π, π ,β₯ (3.29)
where we are using the same notation as the branching ratio. The transduction efficiency scales asπ β |ππ, π ,β₯ππ, π ,β₯|2 if we assume that we are only able to use one of the transition dipole moments (i.e. only electric or only magnetic dipole) for the transduction process. This is the case for us in our standing wave optical cavities and will be mentioned again in the optical cavity design section in Chapter 4.
Figure 3.7: Dipole moment product for a)π1-π1and b)π1-π2for different transition dipole operators (i.e. electric or magnetic) and orientations (i.e. parallel to the a-axis or c-axis).
The dipole product factor forπ1-π1is shown in Figure 3.7a. The strongest transition dipole product forπ1-π1is the magnetic dipole along the c-axis (ππ π·) even though the branching ratio is far from even. Similarly, the dipole product for π1-π2 is shown in Figure 3.7b. Here, the strongest transition dipole for π1-π2is the electric dipole along the c-axis (ππΈ π·), which has both a strong dipole strength and an even branching ratio. Comparing the two different transitions, we would expect βΌ7x larger transduction signal forπ1-π2relative toπ1-π1based on the optical transition dipole strengths and assuming all other parameters are constant.
Initial devices we made were designed to couple to π1-π2 optical transitions, but after some worry about the optical coherence due to the close proximity ofπ1and π2, we switched to devices coupled to π1-π1 optical transitions. Both transitions should be usable for transduction, but our π1-π1 devices had better performance.
This is likely most attributed to more device iterations and improvements, while our π1-π2device was made in the early days of our fabrication process development and did not have the luxury of improvements that were made over the years.
Table 3.5: π1-π1spectroscopic parameters atπ΅π =50Β° Parameter Value
πππ, π1 5.89 πππ,π1 4.55 ππ΅||π,β₯[10β32CΒ·m] 0.48 ππ΅||π,β₯ [10β32 CΒ·m] 3.26 ππ΅||π,β₯ [10β32CΒ·m] 1.58 ππ΅||π,β₯ [10β32CΒ·m] 0.23 ππΈ||π,β₯[10β32 CΒ·m] 0.52 ππΈ||π,β₯ [10β32CΒ·m] 0.86 Best Magnetic Field Orientation for π1-π1
Based on the dipole moment product, for π1-π1 the best magnetic field is at an angle of 35Β° from the c-axis. However, it is also beneficial to have the πππ for the ground and excited state to differ as much as possible which occurs at larger magnetic field angles (see Figure 3.5). Therefore, we decide to use a magnetic field angle of 50Β°as a balance of these two effects. At an angle of 50Β° from the c-axis, the dipole moment product is still near its maximum, so we only expect a decrease in transduction efficiency ofβΌ25%, which is a small factor.
A summary of the magnetic field dependent π1-π1 spectroscopic parameters for a magnetic field angle of 50Β°are summarized in Table 3.5. Theπ1label for the dipole moments was dropped, since we are only including theπ1-π1transitions here.