This work had three objectives to improve an EIT-based, solid-state memory for light. First, we set up a solid-state-laser system for radiation at the wavelength λ = 606nm, i.e., the optical transition in our storage medium, the rare-earth-ion doped crystal PrYSO. Second, we implemented efficient rephasing of optically driven coherences after EIT-based light storage by means of rapid adiabatic passage (RAP) pulses. Last but not least we implemented a novel coherence population mapping (CPM) protocol in order to shelve fragile atomic coherences in robust and long-lived populations in PrYSO. Solid-State-Laser System: We developed a solid-state-laser system based on two nonlinear processes, optical parametric oscillation (OPO) and intra-cavity sum-frequency generation (SFG). The system is designed to generate continuous wave output in the orange part of the visible spectrum. OPO and SFG are implemented on a periodically poled lithium niobate crystal (PPLN). The crystal is divided into sections with appropriate poling periods for quasi phase matching of OPO and SFG. In addition, the poling period changes along the crystal height to allow tuning of the OPO-SFG output wavelength. The system provides output in a range between λ = 605nm and λ = 616nm with an output power P > 1W. For light storage experiments, we operate the OPO-SFG at λ = 606nm with a maximum available output power of P = 1.3W. An external Pound-Drever-Hall (PDH) frequency stabilization reduces the laser linewidth to Δν ≈ 60 kHz on a time scale of 100 ms. The OPO-SFG provides stable output for more than 30 hours with a root-mean-square power jitter below 2%. In comparison to a previously used dye laser, the OPO-SFG requires less maintenance, performs more robust against external temperature fluctuations and vibrational noise and allows easier (re-) alignment. It performs slightly better in terms of frequency stability and the two systems are comparable in terms of output power. Dye lasers on the other hand provide a larger tuning range. Future work should include a revised crystal design, i.e., a reversed section order of OPO and SFG on the PPLN crystal. This has been suggested to yield higher efficiency and output stability. In addition, we use three discrete poling periods in the SFG section, whereas the OPO section consists of a fanned poling structure. A double-fanned structure could provide enhanced tunability and help to partially suppress power fluctuations for different output wavelength. In addition, it should be possible to further reduce the laser-frequency linewidth by revising the locking loop or using a higher finesse cavity for the PDH stabilization. Adiabatic Rephasing of Atomic Coherences: We experimentally implemented rephasing of optically driven coherences in PrYSO by RAP pulses. As a feature of adiabatic pulses, the parameters for RAP are defined in loose boundaries given by the adiabaticity criterion. This makes RAP potentially robust to parameter fluctuations. We experimentally verified this property and showed that rephasing with RAP provides superior performance compared to rephasing with π pulses. In particular, RAP provides enhanced robustness against variations in the Rabi frequency, pulse detuning and spatial inhomogeneity of the driving field. In our specific (3 mm long) PrYSO crystal and for standard experimental parameters, RAP yielded a factor of 1.15 higher rephasing efficiency compared to rephasing with π pulses. This value further increased when we artificially increased experimental imperfections or performed experiments with lower maximum Rabi frequencies. Concluding, RAP provides higher efficiency and more robust performance than π pulses in inhomogeneously broadened media, in the case of low available driving field power and for driving field inhomogeneities. However, RAP pulses typically require longer pulse durations than π pulses. This can cause heating or prevent the use of RAP when fast rephasing is required. An alternative to adiabatic pulses are, e.g., composite pulses or single-shot-shaped pulses. These are investigated in the context of another Ph.D thesis. Coherence Population Mapping: We implemented a novel CPM protocol to store atomic coherences in long-lived populations in PrYSO. CPM works in any three-state system and does not require complex setups beyond a radiation source to generate a short write and read sequence. As an important feature, CPM stores arbitrary coherences equally well in the populations of a three-state system, i.e., CPM does neither require inhomogeneous broadening, nor is the storage efficiency dependent on the phase of the initial coherence. To our best knowledge, this exhibits a unique feature, which no other coherence population mapping protocol provides. Thus, CPM is an alternative to the stimulated photon echo (SPE), which requires inhomogeneous broadening to map an arbitrary initial coherence onto populations. However, the maximum retrieval efficiency with CPM is 1/3 of the initial coherence amplitude (and 1/2 with SPE). We experimentally verified the main characteristics of both protocols with RF driven mapping sequences and RF induced initial coherences. Our results confirm phase-insensitive storage with CPM and storage times reaching the minute regime, i.e., the population relaxation time. We also tested CPM in combination with EIT-based light storage. However, we obtained retrieval efficiencies below 1% and we observed reduced storage durations (for both SPE and CPM) compared to our previous experiments with RF induced coherences. A major contribution to the reduced efficiency is due to the fact that light storage works in a specific ensemble in the inhomogeneously broadened optical line in PrYSO. Our specific CPM (and SPE) pulses on the other hand couple to more ensembles and cause enhanced absorption for the retrieved signal. Future work should thus focus on the implementation of optical CPM to address only the ensemble used for light storage. Regarding a reduced storage duration, we could not yet find an explanation and suggest to implemented CPM in different storage media to further investigate the limitation of storage duration.
This work had three objectives to improve an EIT-based, solid-state memory for light. First, we set up a solid-state-laser system for radiation at the wavelength λ = 606nm, i.e., the optical transition in our storage medium, the rare-earth-ion doped crystal PrYSO. Second, we implemented efficient rephasing of optically driven coherences after EIT-based light storage by means of rapid adiabatic passage (RAP) pulses. Last but not least we implemented a novel coherence population mapping (CPM) protocol in order to shelve fragile atomic coherences in robust and long-lived populations in PrYSO. Solid-State-Laser System: We developed a solid-state-laser system based on two nonlinear processes, optical parametric oscillation (OPO) and intra-cavity sum-frequency generation (SFG). The system is designed to generate continuous wave output in the orange part of the visible spectrum. OPO and SFG are implemented on a periodically poled lithium niobate crystal (PPLN). The crystal is divided into sections with appropriate poling periods for quasi phase matching of OPO and SFG. In addition, the poling period changes along the crystal height to allow tuning of the OPO-SFG output wavelength. The system provides output in a range between λ = 605nm and λ = 616nm with an output power P > 1W. For light storage experiments, we operate the OPO-SFG at λ = 606nm with a maximum available output power of P = 1.3W. An external Pound-Drever-Hall (PDH) frequency stabilization reduces the laser linewidth to Δν ≈ 60 kHz on a time scale of 100 ms. The OPO-SFG provides stable output for more than 30 hours with a root-mean-square power jitter below 2%. In comparison to a previously used dye laser, the OPO-SFG requires less maintenance, performs more robust against external temperature fluctuations and vibrational noise and allows easier (re-) alignment. It performs slightly better in terms of frequency stability and the two systems are comparable in terms of output power. Dye lasers on the other hand provide a larger tuning range. Future work should include a revised crystal design, i.e., a reversed section order of OPO and SFG on the PPLN crystal. This has been suggested to yield higher efficiency and output stability. In addition, we use three discrete poling periods in the SFG section, whereas the OPO section consists of a fanned poling structure. A double-fanned structure could provide enhanced tunability and help to partially suppress power fluctuations for different output wavelength. In addition, it should be possible to further reduce the laser-frequency linewidth by revising the locking loop or using a higher finesse cavity for the PDH stabilization. Adiabatic Rephasing of Atomic Coherences: We experimentally implemented rephasing of optically driven coherences in PrYSO by RAP pulses. As a feature of adiabatic pulses, the parameters for RAP are defined in loose boundaries given by the adiabaticity criterion. This makes RAP potentially robust to parameter fluctuations. We experimentally verified this property and showed that rephasing with RAP provides superior performance compared to rephasing with π pulses. In particular, RAP provides enhanced robustness against variations in the Rabi frequency, pulse detuning and spatial inhomogeneity of the driving field. In our specific (3 mm long) PrYSO crystal and for standard experimental parameters, RAP yielded a factor of 1.15 higher rephasing efficiency compared to rephasing with π pulses. This value further increased when we artificially increased experimental imperfections or performed experiments with lower maximum Rabi frequencies. Concluding, RAP provides higher efficiency and more robust performance than π pulses in inhomogeneously broadened media, in the case of low available driving field power and for driving field inhomogeneities. However, RAP pulses typically require longer pulse durations than π pulses. This can cause heating or prevent the use of RAP when fast rephasing is required. An alternative to adiabatic pulses are, e.g., composite pulses or single-shot-shaped pulses. These are investigated in the context of another Ph.D thesis. Coherence Population Mapping: We implemented a novel CPM protocol to store atomic coherences in long-lived populations in PrYSO. CPM works in any three-state system and does not require complex setups beyond a radiation source to generate a short write and read sequence. As an important feature, CPM stores arbitrary coherences equally well in the populations of a three-state system, i.e., CPM does neither require inhomogeneous broadening, nor is the storage efficiency dependent on the phase of the initial coherence. To our best knowledge, this exhibits a unique feature, which no other coherence population mapping protocol provides. Thus, CPM is an alternative to the stimulated photon echo (SPE), which requires inhomogeneous broadening to map an arbitrary initial coherence onto populations. However, the maximum retrieval efficiency with CPM is 1/3 of the initial coherence amplitude (and 1/2 with SPE). We experimentally verified the main characteristics of both protocols with RF driven mapping sequences and RF induced initial coherences. Our results confirm phase-insensitive storage with CPM and storage times reaching the minute regime, i.e., the population relaxation time. We also tested CPM in combination with EIT-based light storage. However, we obtained retrieval efficiencies below 1% and we observed reduced storage durations (for both SPE and CPM) compared to our previous experiments with RF induced coherences. A major contribution to the reduced efficiency is due to the fact that light storage works in a specific ensemble in the inhomogeneously broadened optical line in PrYSO. Our specific CPM (and SPE) pulses on the other hand couple to more ensembles and cause enhanced absorption for the retrieved signal. Future work should thus focus on the implementation of optical CPM to address only the ensemble used for light storage. Regarding a reduced storage duration, we could not yet find an explanation and suggest to implemented CPM in different storage media to further investigate the limitation of storage duration.
This work had three objectives to improve an EIT-based, solid-state memory for light. First, we set up a solid-state-laser system for radiation at the wavelength λ = 606nm, i.e., the optical transition in our storage medium, the rare-earth-ion doped crystal PrYSO. Second, we implemented efficient rephasing of optically driven coherences after EIT-based light storage by means of rapid adiabatic passage (RAP) pulses. Last but not least we implemented a novel coherence population mapping (CPM) protocol in order to shelve fragile atomic coherences in robust and long-lived populations in PrYSO. Solid-State-Laser System: We developed a solid-state-laser system based on two nonlinear processes, optical parametric oscillation (OPO) and intra-cavity sum-frequency generation (SFG). The system is designed to generate continuous wave output in the orange part of the visible spectrum. OPO and SFG are implemented on a periodically poled lithium niobate crystal (PPLN). The crystal is divided into sections with appropriate poling periods for quasi phase matching of OPO and SFG. In addition, the poling period changes along the crystal height to allow tuning of the OPO-SFG output wavelength. The system provides output in a range between λ = 605nm and λ = 616nm with an output power P > 1W. For light storage experiments, we operate the OPO-SFG at λ = 606nm with a maximum available output power of P = 1.3W. An external Pound-Drever-Hall (PDH) frequency stabilization reduces the laser linewidth to Δν ≈ 60 kHz on a time scale of 100 ms. The OPO-SFG provides stable output for more than 30 hours with a root-mean-square power jitter below 2%. In comparison to a previously used dye laser, the OPO-SFG requires less maintenance, performs more robust against external temperature fluctuations and vibrational noise and allows easier (re-) alignment. It performs slightly better in terms of frequency stability and the two systems are comparable in terms of output power. Dye lasers on the other hand provide a larger tuning range. Future work should include a revised crystal design, i.e., a reversed section order of OPO and SFG on the PPLN crystal. This has been suggested to yield higher efficiency and output stability. In addition, we use three discrete poling periods in the SFG section, whereas the OPO section consists of a fanned poling structure. A double-fanned structure could provide enhanced tunability and help to partially suppress power fluctuations for different output wavelength. In addition, it should be possible to further reduce the laser-frequency linewidth by revising the locking loop or using a higher finesse cavity for the PDH stabilization. Adiabatic Rephasing of Atomic Coherences: We experimentally implemented rephasing of optically driven coherences in PrYSO by RAP pulses. As a feature of adiabatic pulses, the parameters for RAP are defined in loose boundaries given by the adiabaticity criterion. This makes RAP potentially robust to parameter fluctuations. We experimentally verified this property and showed that rephasing with RAP provides superior performance compared to rephasing with π pulses. In particular, RAP provides enhanced robustness against variations in the Rabi frequency, pulse detuning and spatial inhomogeneity of the driving field. In our specific (3 mm long) PrYSO crystal and for standard experimental parameters, RAP yielded a factor of 1.15 higher rephasing efficiency compared to rephasing with π pulses. This value further increased when we artificially increased experimental imperfections or performed experiments with lower maximum Rabi frequencies. Concluding, RAP provides higher efficiency and more robust performance than π pulses in inhomogeneously broadened media, in the case of low available driving field power and for driving field inhomogeneities. However, RAP pulses typically require longer pulse durations than π pulses. This can cause heating or prevent the use of RAP when fast rephasing is required. An alternative to adiabatic pulses are, e.g., composite pulses or single-shot-shaped pulses. These are investigated in the context of another Ph.D thesis. Coherence Population Mapping: We implemented a novel CPM protocol to store atomic coherences in long-lived populations in PrYSO. CPM works in any three-state system and does not require complex setups beyond a radiation source to generate a short write and read sequence. As an important feature, CPM stores arbitrary coherences equally well in the populations of a three-state system, i.e., CPM does neither require inhomogeneous broadening, nor is the storage efficiency dependent on the phase of the initial coherence. To our best knowledge, this exhibits a unique feature, which no other coherence population mapping protocol provides. Thus, CPM is an alternative to the stimulated photon echo (SPE), which requires inhomogeneous broadening to map an arbitrary initial coherence onto populations. However, the maximum retrieval efficiency with CPM is 1/3 of the initial coherence amplitude (and 1/2 with SPE). We experimentally verified the main characteristics of both protocols with RF driven mapping sequences and RF induced initial coherences. Our results confirm phase-insensitive storage with CPM and storage times reaching the minute regime, i.e., the population relaxation time. We also tested CPM in combination with EIT-based light storage. However, we obtained retrieval efficiencies below 1% and we observed reduced storage durations (for both SPE and CPM) compared to our previous experiments with RF induced coherences. A major contribution to the reduced efficiency is due to the fact that light storage works in a specific ensemble in the inhomogeneously broadened optical line in PrYSO. Our specific CPM (and SPE) pulses on the other hand couple to more ensembles and cause enhanced absorption for the retrieved signal. Future work should thus focus on the implementation of optical CPM to address only the ensemble used for light storage. Regarding a reduced storage duration, we could not yet find an explanation and suggest to implemented CPM in different storage media to further investigate the limitation of storage duration.
The present work deals with techniques to improve the EIT-based light storage in an all solid-state memory, i.e. a rare-earth ion-doped PrYSO crystal. The performance of such a memory can be defined by its capability to store light pulses with high efficiency and for long storage durations. In general, the storage efficiency is theoretically limited by the EIT-LS protocol efficiency, while the storage duration is limited by decoherence processes in the solid-state memory. Thus, we first optimized the EIT-LS efficiency at short storage durations. We then investigated the performance of composite and adiabatic techniques for efficient and robust rephasing and finally applied composite techniques for DDC to extend the coherence lifetime in our storage medium. Optimization of the EIT Light Storage Efficiency: In order to optimize the EIT-LS protocol we performed systematic measurements of the EIT-LS efficiency in a PrYSO crystal featuring an optical depth d of 6. The optical depth sets the theoretical limit of the EIT-LS protocol efficiency. We experimentally optimized the efficiency by systematic variations of the control pulse power and the probe pulse duration. Furthermore, we applied an iterative algorithm, to optimize the temporal shape of the probe pulse. We found a maximal protocol efficiency 36 %, comparable to the theoretical limit. In order to increase the optical depth d, we developed a multipass setup for the probe beam, consisting of a ringlike arrangement, which is compatible with the geometrical constraints given by the existing setup. This multipass setup allows the simple variation of the number of probe passes N through the crystal and thus enables a flexible change of the effective optical depth. With this setup we achieved up to 16 probe passes through the PrYSO crystal, which corresponds to an increase of the effective optical depth from 6 to 96. We experimentally optimized the EIT-LS efficiency at variable optical depths. At N=14, i.e. an effective optical depth of about 84, we achieved an EIT-LS efficiency of 76.3 % in forward retrieval configuration, reaching previous values in an EIT-driven memory of cold atoms and achieving the highest ever obtained EIT-LS efficiency in a solid-state memory. However, due to losses at the optical components, the setup efficiency was limited to about 25.2 % at N=2. As future work will focus on the storage of few photons, it will thus be necessary to improve the optical components, to achieve adequate detection efficiencies. It might also be useful, to further investigate possibilities for backward readout configurations combined with a multipass probe setup. Composite and Adiabatic Rephasing of Atomic Coherences: In order to extend the EIT-LS duration of our memory towards the coherence lifetime, we implemented composite and adiabatic rephasing techniques, exhibiting an improved robustness regarding variations and fluctuations of experimental parameters. We investigated their performance regarding experimental variations, in an application for rephasing of atomic coherences in the inhomogeneously broadened hyperfine transition of PrYSO. First, we applied universal composite pulses (UCP). UCP were originally designed for robust and high-fidelity population inversion in a two-state system, compensating simultaneous variations in any type of experimental parameter. We have experimentally shown that UCP exhibit an enhanced robustness to variations in several experimental parameters. Using UCP we could increase the higher rephasing efficiency about 25 %, compared to diabatic π pulses. UCP can be useful whenever significant unknown experimental variations or fluctuations prevent the application of diabatic π pulses. Second, we demonstrated a first experimental implementation of composite adiabatic passage (CAP). Essentially, CAP is a composite version of RAP, consisting of a sequence of RAP pulses with appropriately chosen relative phases. CAP was proposed to improve RAP in situations of fairly fulfilled adiabaticity. We compared the performance of CAP and RAP at different degrees of adiabaticity. In particular, we systematically investigated their performance with respect to variations in pulse duration and static detuning. We found CAP to be able to compensate for weak adiabaticity, leading to constant and high rephasing efficiency, irrespectively of the exact choice of the experimental parameters. These properties of CAP can be of interest whenever robust state manipulations are required, while sufficient adiabaticity cannot be provided, e.g. due to experimental restrictions. Third, we performed a first experimental demonstration of single-shot shaped pulses (SSSP), derived from techniques on shortcuts to adiabaticity. We demonstrated the capability of SSSP for efficient and robust rephasing and compared our results with diabatic π pulses. A further comparison with other adiabatic techniques might be necessary to provide a broader insight on SSSP. All three presented techniques can be used to improve the robustness of an population inversion or rephasing process. However, different requirements on the control of the experimental parameters have to be fulfilled. UCP, and CAP rely on identical pulses and a precise relative phase control, i.e. within a few degrees. CAP in addition needs a simple control of the time-dependent detuning. The complex time-dependent Rabi frequency and detuning of SSSP require a much more sophisticate control of experimental parameters. Thus, the choice of an adequate technique strongly depends on the actual experimental situation, i.e. on the experimentally controllable parameters. Composite Pulses for Dynamic Decoherence Control: We investigated the performance of UCP and universal robust (UR) sequences for DDC of directly RF-prepared coherences. We compared our results with well known CPMG DD and KDD sequences. We performed systematic measurements with respect to the cycling time, the phase of the coherence and the order of UCP and UR DD sequence. We found time separated UR DD sequences to be robust with respect to the phase of the coherence. We applied these DD sequences on EIT-LS coherences and compared with CPMG DD and KDD in XY4. Our experiments showed that our UR DD sequences can outperform the often considered state-of-the-art KDD in XY4 sequence, yielding about a factor of 2 longer coherence lifetimes. Combining our results on optimized EIT-LS with advanced composite DD sequences and static decoherence control by ZEFOZ might provide a major step towards the development of an all solid-state quantum memory.
The present work deals with techniques to improve the EIT-based light storage in an all solid-state memory, i.e. a rare-earth ion-doped PrYSO crystal. The performance of such a memory can be defined by its capability to store light pulses with high efficiency and for long storage durations. In general, the storage efficiency is theoretically limited by the EIT-LS protocol efficiency, while the storage duration is limited by decoherence processes in the solid-state memory. Thus, we first optimized the EIT-LS efficiency at short storage durations. We then investigated the performance of composite and adiabatic techniques for efficient and robust rephasing and finally applied composite techniques for DDC to extend the coherence lifetime in our storage medium. Optimization of the EIT Light Storage Efficiency: In order to optimize the EIT-LS protocol we performed systematic measurements of the EIT-LS efficiency in a PrYSO crystal featuring an optical depth d of 6. The optical depth sets the theoretical limit of the EIT-LS protocol efficiency. We experimentally optimized the efficiency by systematic variations of the control pulse power and the probe pulse duration. Furthermore, we applied an iterative algorithm, to optimize the temporal shape of the probe pulse. We found a maximal protocol efficiency 36 %, comparable to the theoretical limit. In order to increase the optical depth d, we developed a multipass setup for the probe beam, consisting of a ringlike arrangement, which is compatible with the geometrical constraints given by the existing setup. This multipass setup allows the simple variation of the number of probe passes N through the crystal and thus enables a flexible change of the effective optical depth. With this setup we achieved up to 16 probe passes through the PrYSO crystal, which corresponds to an increase of the effective optical depth from 6 to 96. We experimentally optimized the EIT-LS efficiency at variable optical depths. At N=14, i.e. an effective optical depth of about 84, we achieved an EIT-LS efficiency of 76.3 % in forward retrieval configuration, reaching previous values in an EIT-driven memory of cold atoms and achieving the highest ever obtained EIT-LS efficiency in a solid-state memory. However, due to losses at the optical components, the setup efficiency was limited to about 25.2 % at N=2. As future work will focus on the storage of few photons, it will thus be necessary to improve the optical components, to achieve adequate detection efficiencies. It might also be useful, to further investigate possibilities for backward readout configurations combined with a multipass probe setup. Composite and Adiabatic Rephasing of Atomic Coherences: In order to extend the EIT-LS duration of our memory towards the coherence lifetime, we implemented composite and adiabatic rephasing techniques, exhibiting an improved robustness regarding variations and fluctuations of experimental parameters. We investigated their performance regarding experimental variations, in an application for rephasing of atomic coherences in the inhomogeneously broadened hyperfine transition of PrYSO. First, we applied universal composite pulses (UCP). UCP were originally designed for robust and high-fidelity population inversion in a two-state system, compensating simultaneous variations in any type of experimental parameter. We have experimentally shown that UCP exhibit an enhanced robustness to variations in several experimental parameters. Using UCP we could increase the higher rephasing efficiency about 25 %, compared to diabatic π pulses. UCP can be useful whenever significant unknown experimental variations or fluctuations prevent the application of diabatic π pulses. Second, we demonstrated a first experimental implementation of composite adiabatic passage (CAP). Essentially, CAP is a composite version of RAP, consisting of a sequence of RAP pulses with appropriately chosen relative phases. CAP was proposed to improve RAP in situations of fairly fulfilled adiabaticity. We compared the performance of CAP and RAP at different degrees of adiabaticity. In particular, we systematically investigated their performance with respect to variations in pulse duration and static detuning. We found CAP to be able to compensate for weak adiabaticity, leading to constant and high rephasing efficiency, irrespectively of the exact choice of the experimental parameters. These properties of CAP can be of interest whenever robust state manipulations are required, while sufficient adiabaticity cannot be provided, e.g. due to experimental restrictions. Third, we performed a first experimental demonstration of single-shot shaped pulses (SSSP), derived from techniques on shortcuts to adiabaticity. We demonstrated the capability of SSSP for efficient and robust rephasing and compared our results with diabatic π pulses. A further comparison with other adiabatic techniques might be necessary to provide a broader insight on SSSP. All three presented techniques can be used to improve the robustness of an population inversion or rephasing process. However, different requirements on the control of the experimental parameters have to be fulfilled. UCP, and CAP rely on identical pulses and a precise relative phase control, i.e. within a few degrees. CAP in addition needs a simple control of the time-dependent detuning. The complex time-dependent Rabi frequency and detuning of SSSP require a much more sophisticate control of experimental parameters. Thus, the choice of an adequate technique strongly depends on the actual experimental situation, i.e. on the experimentally controllable parameters. Composite Pulses for Dynamic Decoherence Control: We investigated the performance of UCP and universal robust (UR) sequences for DDC of directly RF-prepared coherences. We compared our results with well known CPMG DD and KDD sequences. We performed systematic measurements with respect to the cycling time, the phase of the coherence and the order of UCP and UR DD sequence. We found time separated UR DD sequences to be robust with respect to the phase of the coherence. We applied these DD sequences on EIT-LS coherences and compared with CPMG DD and KDD in XY4. Our experiments showed that our UR DD sequences can outperform the often considered state-of-the-art KDD in XY4 sequence, yielding about a factor of 2 longer coherence lifetimes. Combining our results on optimized EIT-LS with advanced composite DD sequences and static decoherence control by ZEFOZ might provide a major step towards the development of an all solid-state quantum memory.
The research project dealt with the implementation of EIT light storage in the solid state medium Pr:YSO for storage of single photons on a microsecond timescale and weak coherent pulses on a timescale of seconds. The main challenge was to maintain a large efficiency at long storage times and suppress control pulse background by about 120 dB to enable detection of signals on the single photon level. We used a configuration with counterpropagating probe and control beams which enabled a control pulse suppression of about 80 dB and achieved the remaining suppression by spectral filtering. Since the signal and control read pulse frequencies differ only by ∼10 MHz, we built a filter based on a second Pr:YSO crystal in which we tailored the absorption spectrum to the experimental requirements by optical pumping. For short storage times of 2 μs, the filter enhanced the separation between the signal and control read pulses by 42 dB. This enabled the storage of a weak coherent pulse with an average photon number of 1.1 with an efficiency of 42% and a SNR of 3.2. The efficiency was limited by the OD of the storage medium and could be increased by a multipass setup. Residual leakage of the control pulse through the filter limited the SNR. Still, to the best of our knowledge, our result represents the first implementation of single photon storage based on EIT in a solid state medium. To prolong the coherence lifetime, we implemented ZEFOZ. The influence of the magnetic field reduced the OD of the Pr:YSO crystal which led to a lower storage efficiency. We characterized light storage under ZEFOZ conditions at a short storage time of 2 μs. By application of a multipass setup with three passes and temporal shaping of the probe pulse, we partially compensated the reduced OD and achieved efficiencies of up to 23% for storage of classical light pulses. In order to maintain a high setup transmission, we used only two passes for storage of weak coherent pulses. We demonstrated storage of a weak coherent pulse with an average of 10.2 photons for 2 μs with 16% efficiency and a SNR of 1.3. Again, leakage of the control read pulse through the spectral filter limited the SNR. We found this leakage to be due to a weak but spectrally broad pedestal in the spectrum of our OPO-SFG laser system. In order to prolong the light storage time and exploit the long coherence lifetime under ZEFOZ conditions, we applied different sequences for rephasing with RF pulses. In the design of our RF coils for rephasing, we had to consider restrictions given by the static magnetic field setup. Therefore, the rephasing π pulses had a significant inhomogeneity which led to a low rephasing efficiency of 15% when we applied two π pulses. To compensate these errors, we applied UR DD which increased the rephasing efficiency to 40%. Due to an experimental limitation, we could not reduce the separation of the rephasing π pulses below 10 ms without reducing the storage efficiency. This limited the effectiveness of DD. Nevertheless, we achieved a storage time of 7.5 s for classical light pulses and confirmed this timescale for storage of weak coherent pulses with an average of 52 photons. We stored weak coherent pulses with different photon number for 1.28 s and extrapolated a SNR of 0.09 for storage of a single photon. For storage of a weak coherent pulse with an average of 7 photons, we determined an efficiency of 6% and a SNR of 0.6. This represents the first implementation of a memory for weak coherent pulses close to the single photon level with a storage time in the regime of seconds. Compared to the current state-of-the-art single photon storage time of 100 ms, we achieve a more than one order of magnitude longer storage time. As an extension of our investigations, we developed, implemented and studied novel universal composite pulse sequences to invert superposition states or efficiently transfer population between two states. In particular, we experimentally demonstrated UCPs for population inversion which are especially suited for situations in which pulse area and detuning errors are correlated. We showed that we can rotate the excitation profile of the sequence by changing the phases of the excitation pulses. This enables adaption to the specific correlation of errors present in an experiment. Moreover, we demonstrated a composite version of STIRAP. We showed that for detuned pulses, the phases of UCPs enhance the robustness to variations in any experimental parameter. We achieved an increase in population transfer efficiency from 50% for five times repeated detuned STIRAP to 85% for UR CSTIRAP.
This thesis describes experiments aimed at developing a key component in applications such as quantum networks and long-distance quantum communication systems: quantum memories. Quantum memories for light using photon-echo and coherent rephrasing techniques have been developed. The work is carried out on rare-earth-ion-doped inorganic crystals. Several properties are associated with these crystals such as long coherence times and long-lived ground state sublevels that are particularly interesting in connection applications in the area of quantum information science. It is demonstrated the preparation of optical inhomogeneous absorption profiles such that the light only interacts with a specific transition at a selected frequency. Narrow absorbing structures with widths of <100 kHz have been prepared with no absorption in the surrounding spectral interval. Most of the experiments were carried out using Pr3+:Y2SiO5. However, in our search for suitable a Λ-system in the material under consideration, we also performed spectral hole-burning spectroscopy in Nd:Y2SiVO4. As part of the search for host materials for Pr, Pr:La2(WO4)3 crystal was characterized using a variety of methods, such as, hole burning and photon-echoes. Electromagnetically induced transparency was also observed in this crystal. The most important criteria for evaluating quantum memory performance and the most important physical requirements are discussed. After investigating the standard photon echo techniques and controlled reversible inhomogeneous broadening (CRIB) for quantum state storage, we considered an efficient multimode quantum memory protocol; the atomic frequency comb (AFC) protocol. Combs with peaks of widths 100-300 kHz and an optical depth of~ 10 were created by optical pumping inside an emptied region of the inhomogeneous profile of Pr3+:Y2SiO5, allowing us to improve the efficiency of light storage. Light storage combining a photon-echo technique based on an AFC and spin-wave storage have been demonstrated, and an experimental AFC scheme for the storage of weak coherent light pulses in Pr3+:Y2SiO5 is presented. Finally, we have investigated superradiance and slow light effects. Both these effects occur in the high optical depth regime and can influence the memory performance. Under certain conditions superradiance results in the immediate re-emission of the stored light in a single burst of coherent radiation, and inside spectrally structured materials, slow light effects may result when storage is performed.
Efficient and long-lived multimode quantum memories are crucial devices in the development of quantum technolgies. The reversible mapping of quantum states of light in rare earth doped crystals represents one of the most promising routes towards the realization of this goal. Such systems are also compatible with the miniaturization of quantum memories in integrated optics platforms, which offer unique features in terms of experimental scalability and enhanced light-matter interaction. Here, we fabricate single mode channel waveguides for 606 nm light in a praseodymium-doped yttrium orthosilicate crystal (Pr3+:Y2SiO5), that, thanks to its excellent coherence properties, is a widely studied material for light storage experiments. Waveguides are inscribed by femtosecond laser writing, adopting the so-called Type I configuration, where the core is directly obtained at the irradiated area. Remarkably, fabricating this kind of waveguides in crystals is a difficult task, as it requires to operate in a very narrow processing parameters window, if existing. We then use these waveguides for performing the storage and retrieval of single photons, implementing the atomic frequency comb protocol. We achieve a storage time of 5,5 µs, which is almost 2 orders of magnitude longer than previous realizations of quantum light storage in a waveguide. In addition, we investigate the potential information multiplexing capabilities of our system by performing the quantum storage of single photons delocalized over 14 different spectral modes. Our results show that laser written waveguides in rare earth-doped solid state systems are very promising for the development of efficient and long-lived multimode quantum memories.