The proton polarization in the 3He(d,p)4He reaction induced by unpolarized deuterons has been measured at deuteron lab. energies of 2.0, 2.8, 3.9 and 6.0MeV for 20 angles between 0° and 150° (c.m.). Statistical uncertainties are typically ± 0.01. The measurements were made with a proton polarimeter in which the left-right asymmetry of scattering at 60° (lab.) in 4He is determined. The polarimeter employs "venetian-blind" collimation of the protons by conical vanes and 75 cm2 plastic scintillator detectors. Four detectors are included for use in polarization transfer experiments. For 10.5 MeV protons and a helium pressure of 250 p.s.i. the target thickness is 3 MeV and the efficiency per detector per unpolarized proton incident is 10-4. For each polarimeter detector a triple coincidence with a 15 ns resolving time was required with two scintillator transmission detectors preceding the polarimeter. Spectra of random coincidences were accumulated simultaneously and subtracted. Asymmetries resulting from polarimeter-target misalignment and other geometrical effects are discussed. All results quoted are geometric means of pairs of measurements for 180° rotation of the polarimeter and are also arithmetic means of such measurements to left and right of the 3He target. The absolute analyzing power is estimated by computer simulation of trajectories to be -0.638 ± 0.020 for protons entering at 10.3 MeV. The product of polarization and cross section is fitted to an expansion of first-order associated Legendre polynomials using these results and earlier measurements. Only four terms are required except at 6.0MeV where a fifth is necessary. The energy dependence of these coefficients suggests resonances in 5Li at deuteron energies of 60MeV (odd coefficients) and 7.5 MeV (even coefficients) in agreement with results for the polarized-beam analyzing powers(1). Comparison of the results with vector-polarized-beam (1) and polarized-target(2) analyzing powers shows no evidence for the postulated simple relations(3) based on DWBA cal calculations. Comparison of the results with recent measurements of the neutron polarization in the mirror reaction(4) shows no significant differences. The theory of angular correlations in charged particle reactions is developed and used to calculate outgoing nucleon polarizations. Expressions are given for polarization transfer coefficients. These coefficients are evaluated in terms of the T-matrix elements for the interference of various channels with the dominant S-wave, JΠ = 3+/2 channel in 3He(d,p)4He at the 0.43 MeV resonance. Two experiments to measure combinations of these elements are discussed. (1) Gruebler, W. et al., 1971, Nucl. Phys. Al76, 631 (2) Leemann, Ch., W. Gruebler et al., 1971, in Polarization Phenomena in Nuclear Reactions (University of Wisconsin Press), p. 548 (3) Tanifuji,M. and K. Yazaki, 1968, Prog. Theor. Phys. 40, 1023 (4) Mutchler, G.S., W.B. Broste and J.E. Simmons, 1971, Phys. Rev. C3, 1031

This thesis investigates the properties of an atomic system exhibiting a giant Kerr nonlinearity. The atomic energy level scheme involves four energy levels. A three level A subsystem in the atom exhibits the effect of electromagnetically induced transparency (EIT), reducing the spontaneous emission noise. The fourth level leads to an ac-Stark shift of the ground state, which in turn leads to a giant, noiseless Kerr nonlinearity. Two different environments are explored. First, a system comprising of large number of atoms in an optical cavity is analysed. Detailed aspects of noise reduction in this system are investigated. In particular, strong squeezing in the quadrature in phase with the field driving the cavity mode is found, if the effective coupling of light to the atoms is strong. However, the linewidth of the predicted squeezing is found to be very narrow. This is attributed to a very steep linear susceptibility of the atomic medium. Since the widening of the squeezing window is possible only for weaker effective coupling, in turn reducing the squeezing level, a different environment is proposed. This involves a single four level atom, strongly coupled to the cavity mode. In such a strongly coupled system, the most appropriate approach is found to be that formulated in terms of polaritons – composite excitations of the 'atom-cavity molecule'. Adopting the polariton approach, nonclassical correlations in the field leaving the cavity are investigated. Strong photon antibunching is found and the effect of photon blockade predicted and described. The photon blockade effect can also be found in a system comprised of a two level atom coupled to the cavity mode, if the external driving is tuned to one of the vacuum Rabi resonances. A comparison between the two schemes is performed, and it is found that the four level scheme exhibits much better photon blockade. The reason for this is quantum interference between secondary transitions in the dressed states picture. Destructive interference cancels the transitions that would otherwise introduce a second photon into the system, hence producing a more robust photon blockade. All of these results are valid in the regime where external driving is weak. If the external driving strength is increased, the photon statistics (as measured by the zero-delay second order correlation function) changes from strong antibunching to strong bunching, over a relatively narrow range of driving strengths. The occurrence of this change can again be attributed to quantum interference. It is shown that the interference effect prevents the excitation of the composite system by a second photon, but not excitation by a two-photon transition (following the first excitation). Therefore, the third excitation manifold is excited, which then decays back to the first manifold in a two photon cascade. This two photon cascade is the source of correlated photon pairs causing an increase in the second order correlation function. The dynamics of forward scattering of light is presented, and nonclassical behaviour of the delay dependence of correlation function ('overshoots' and 'undershoots') is discussed. For the analytical treatment of this system, a method based on the polariton approach is devised, which includes the treatment of driving and damping. It is shown that this method is ideally suited to the analysis of strongly coupled systems, where only a few photons contribute to the dynamics.

The proton polarization in the 3He(d,p)4He reaction induced by unpolarized deuterons has been measured at deuteron lab. energies of 2.0, 2.8, 3.9 and 6.0MeV for 20 angles between 0° and 150° (c.m.). Statistical uncertainties are typically ± 0.01. The measurements were made with a proton polarimeter in which the left-right asymmetry of scattering at 60° (lab.) in 4He is determined. The polarimeter employs "venetian-blind" collimation of the protons by conical vanes and 75 cm2 plastic scintillator detectors. Four detectors are included for use in polarization transfer experiments. For 10.5 MeV protons and a helium pressure of 250 p.s.i. the target thickness is 3 MeV and the efficiency per detector per unpolarized proton incident is 10-4. For each polarimeter detector a triple coincidence with a 15 ns resolving time was required with two scintillator transmission detectors preceding the polarimeter. Spectra of random coincidences were accumulated simultaneously and subtracted. Asymmetries resulting from polarimeter-target misalignment and other geometrical effects are discussed. All results quoted are geometric means of pairs of measurements for 180° rotation of the polarimeter and are also arithmetic means of such measurements to left and right of the 3He target. The absolute analyzing power is estimated by computer simulation of trajectories to be -0.638 ± 0.020 for protons entering at 10.3 MeV. The product of polarization and cross section is fitted to an expansion of first-order associated Legendre polynomials using these results and earlier measurements. Only four terms are required except at 6.0MeV where a fifth is necessary. The energy dependence of these coefficients suggests resonances in 5Li at deuteron energies of 60MeV (odd coefficients) and 7.5 MeV (even coefficients) in agreement with results for the polarized-beam analyzing powers(1). Comparison of the results with vector-polarized-beam (1) and polarized-target(2) analyzing powers shows no evidence for the postulated simple relations(3) based on DWBA cal calculations. Comparison of the results with recent measurements of the neutron polarization in the mirror reaction(4) shows no significant differences. The theory of angular correlations in charged particle reactions is developed and used to calculate outgoing nucleon polarizations. Expressions are given for polarization transfer coefficients. These coefficients are evaluated in terms of the T-matrix elements for the interference of various channels with the dominant S-wave, JΠ = 3+/2 channel in 3He(d,p)4He at the 0.43 MeV resonance. Two experiments to measure combinations of these elements are discussed. (1) Gruebler, W. et al., 1971, Nucl. Phys. Al76, 631 (2) Leemann, Ch., W. Gruebler et al., 1971, in Polarization Phenomena in Nuclear Reactions (University of Wisconsin Press), p. 548 (3) Tanifuji,M. and K. Yazaki, 1968, Prog. Theor. Phys. 40, 1023 (4) Mutchler, G.S., W.B. Broste and J.E. Simmons, 1971, Phys. Rev. C3, 1031

This thesis investigates the properties of an atomic system exhibiting a giant Kerr nonlinearity. The atomic energy level scheme involves four energy levels. A three level A subsystem in the atom exhibits the effect of electromagnetically induced transparency (EIT), reducing the spontaneous emission noise. The fourth level leads to an ac-Stark shift of the ground state, which in turn leads to a giant, noiseless Kerr nonlinearity. Two different environments are explored. First, a system comprising of large number of atoms in an optical cavity is analysed. Detailed aspects of noise reduction in this system are investigated. In particular, strong squeezing in the quadrature in phase with the field driving the cavity mode is found, if the effective coupling of light to the atoms is strong. However, the linewidth of the predicted squeezing is found to be very narrow. This is attributed to a very steep linear susceptibility of the atomic medium. Since the widening of the squeezing window is possible only for weaker effective coupling, in turn reducing the squeezing level, a different environment is proposed. This involves a single four level atom, strongly coupled to the cavity mode. In such a strongly coupled system, the most appropriate approach is found to be that formulated in terms of polaritons – composite excitations of the 'atom-cavity molecule'. Adopting the polariton approach, nonclassical correlations in the field leaving the cavity are investigated. Strong photon antibunching is found and the effect of photon blockade predicted and described. The photon blockade effect can also be found in a system comprised of a two level atom coupled to the cavity mode, if the external driving is tuned to one of the vacuum Rabi resonances. A comparison between the two schemes is performed, and it is found that the four level scheme exhibits much better photon blockade. The reason for this is quantum interference between secondary transitions in the dressed states picture. Destructive interference cancels the transitions that would otherwise introduce a second photon into the system, hence producing a more robust photon blockade. All of these results are valid in the regime where external driving is weak. If the external driving strength is increased, the photon statistics (as measured by the zero-delay second order correlation function) changes from strong antibunching to strong bunching, over a relatively narrow range of driving strengths. The occurrence of this change can again be attributed to quantum interference. It is shown that the interference effect prevents the excitation of the composite system by a second photon, but not excitation by a two-photon transition (following the first excitation). Therefore, the third excitation manifold is excited, which then decays back to the first manifold in a two photon cascade. This two photon cascade is the source of correlated photon pairs causing an increase in the second order correlation function. The dynamics of forward scattering of light is presented, and nonclassical behaviour of the delay dependence of correlation function ('overshoots' and 'undershoots') is discussed. For the analytical treatment of this system, a method based on the polariton approach is devised, which includes the treatment of driving and damping. It is shown that this method is ideally suited to the analysis of strongly coupled systems, where only a few photons contribute to the dynamics.

The proton polarization in the 3He(d,p)4He reaction induced by unpolarized deuterons has been measured at deuteron lab. energies of 2.0, 2.8, 3.9 and 6.0MeV for 20 angles between 0° and 150° (c.m.). Statistical uncertainties are typically ± 0.01. The measurements were made with a proton polarimeter in which the left-right asymmetry of scattering at 60° (lab.) in 4He is determined. The polarimeter employs "venetian-blind" collimation of the protons by conical vanes and 75 cm2 plastic scintillator detectors. Four detectors are included for use in polarization transfer experiments. For 10.5 MeV protons and a helium pressure of 250 p.s.i. the target thickness is 3 MeV and the efficiency per detector per unpolarized proton incident is 10-4. For each polarimeter detector a triple coincidence with a 15 ns resolving time was required with two scintillator transmission detectors preceding the polarimeter. Spectra of random coincidences were accumulated simultaneously and subtracted. Asymmetries resulting from polarimeter-target misalignment and other geometrical effects are discussed. All results quoted are geometric means of pairs of measurements for 180° rotation of the polarimeter and are also arithmetic means of such measurements to left and right of the 3He target. The absolute analyzing power is estimated by computer simulation of trajectories to be -0.638 ± 0.020 for protons entering at 10.3 MeV. The product of polarization and cross section is fitted to an expansion of first-order associated Legendre polynomials using these results and earlier measurements. Only four terms are required except at 6.0MeV where a fifth is necessary. The energy dependence of these coefficients suggests resonances in 5Li at deuteron energies of 60MeV (odd coefficients) and 7.5 MeV (even coefficients) in agreement with results for the polarized-beam analyzing powers(1). Comparison of the results with vector-polarized-beam (1) and polarized-target(2) analyzing powers shows no evidence for the postulated simple relations(3) based on DWBA cal calculations. Comparison of the results with recent measurements of the neutron polarization in the mirror reaction(4) shows no significant differences. The theory of angular correlations in charged particle reactions is developed and used to calculate outgoing nucleon polarizations. Expressions are given for polarization transfer coefficients. These coefficients are evaluated in terms of the T-matrix elements for the interference of various channels with the dominant S-wave, JΠ = 3+/2 channel in 3He(d,p)4He at the 0.43 MeV resonance. Two experiments to measure combinations of these elements are discussed. (1) Gruebler, W. et al., 1971, Nucl. Phys. Al76, 631 (2) Leemann, Ch., W. Gruebler et al., 1971, in Polarization Phenomena in Nuclear Reactions (University of Wisconsin Press), p. 548 (3) Tanifuji,M. and K. Yazaki, 1968, Prog. Theor. Phys. 40, 1023 (4) Mutchler, G.S., W.B. Broste and J.E. Simmons, 1971, Phys. Rev. C3, 1031

This thesis investigates the properties of an atomic system exhibiting a giant Kerr nonlinearity. The atomic energy level scheme involves four energy levels. A three level A subsystem in the atom exhibits the effect of electromagnetically induced transparency (EIT), reducing the spontaneous emission noise. The fourth level leads to an ac-Stark shift of the ground state, which in turn leads to a giant, noiseless Kerr nonlinearity. Two different environments are explored. First, a system comprising of large number of atoms in an optical cavity is analysed. Detailed aspects of noise reduction in this system are investigated. In particular, strong squeezing in the quadrature in phase with the field driving the cavity mode is found, if the effective coupling of light to the atoms is strong. However, the linewidth of the predicted squeezing is found to be very narrow. This is attributed to a very steep linear susceptibility of the atomic medium. Since the widening of the squeezing window is possible only for weaker effective coupling, in turn reducing the squeezing level, a different environment is proposed. This involves a single four level atom, strongly coupled to the cavity mode. In such a strongly coupled system, the most appropriate approach is found to be that formulated in terms of polaritons – composite excitations of the 'atom-cavity molecule'. Adopting the polariton approach, nonclassical correlations in the field leaving the cavity are investigated. Strong photon antibunching is found and the effect of photon blockade predicted and described. The photon blockade effect can also be found in a system comprised of a two level atom coupled to the cavity mode, if the external driving is tuned to one of the vacuum Rabi resonances. A comparison between the two schemes is performed, and it is found that the four level scheme exhibits much better photon blockade. The reason for this is quantum interference between secondary transitions in the dressed states picture. Destructive interference cancels the transitions that would otherwise introduce a second photon into the system, hence producing a more robust photon blockade. All of these results are valid in the regime where external driving is weak. If the external driving strength is increased, the photon statistics (as measured by the zero-delay second order correlation function) changes from strong antibunching to strong bunching, over a relatively narrow range of driving strengths. The occurrence of this change can again be attributed to quantum interference. It is shown that the interference effect prevents the excitation of the composite system by a second photon, but not excitation by a two-photon transition (following the first excitation). Therefore, the third excitation manifold is excited, which then decays back to the first manifold in a two photon cascade. This two photon cascade is the source of correlated photon pairs causing an increase in the second order correlation function. The dynamics of forward scattering of light is presented, and nonclassical behaviour of the delay dependence of correlation function ('overshoots' and 'undershoots') is discussed. For the analytical treatment of this system, a method based on the polariton approach is devised, which includes the treatment of driving and damping. It is shown that this method is ideally suited to the analysis of strongly coupled systems, where only a few photons contribute to the dynamics.

The proton polarization in the 3He(d,p)4He reaction induced by unpolarized deuterons has been measured at deuteron lab. energies of 2.0, 2.8, 3.9 and 6.0MeV for 20 angles between 0° and 150° (c.m.). Statistical uncertainties are typically ± 0.01. The measurements were made with a proton polarimeter in which the left-right asymmetry of scattering at 60° (lab.) in 4He is determined. The polarimeter employs "venetian-blind" collimation of the protons by conical vanes and 75 cm2 plastic scintillator detectors. Four detectors are included for use in polarization transfer experiments. For 10.5 MeV protons and a helium pressure of 250 p.s.i. the target thickness is 3 MeV and the efficiency per detector per unpolarized proton incident is 10-4. For each polarimeter detector a triple coincidence with a 15 ns resolving time was required with two scintillator transmission detectors preceding the polarimeter. Spectra of random coincidences were accumulated simultaneously and subtracted. Asymmetries resulting from polarimeter-target misalignment and other geometrical effects are discussed. All results quoted are geometric means of pairs of measurements for 180° rotation of the polarimeter and are also arithmetic means of such measurements to left and right of the 3He target. The absolute analyzing power is estimated by computer simulation of trajectories to be -0.638 ± 0.020 for protons entering at 10.3 MeV. The product of polarization and cross section is fitted to an expansion of first-order associated Legendre polynomials using these results and earlier measurements. Only four terms are required except at 6.0MeV where a fifth is necessary. The energy dependence of these coefficients suggests resonances in 5Li at deuteron energies of 60MeV (odd coefficients) and 7.5 MeV (even coefficients) in agreement with results for the polarized-beam analyzing powers(1). Comparison of the results with vector-polarized-beam (1) and polarized-target(2) analyzing powers shows no evidence for the postulated simple relations(3) based on DWBA cal calculations. Comparison of the results with recent measurements of the neutron polarization in the mirror reaction(4) shows no significant differences. The theory of angular correlations in charged particle reactions is developed and used to calculate outgoing nucleon polarizations. Expressions are given for polarization transfer coefficients. These coefficients are evaluated in terms of the T-matrix elements for the interference of various channels with the dominant S-wave, JΠ = 3+/2 channel in 3He(d,p)4He at the 0.43 MeV resonance. Two experiments to measure combinations of these elements are discussed. (1) Gruebler, W. et al., 1971, Nucl. Phys. Al76, 631 (2) Leemann, Ch., W. Gruebler et al., 1971, in Polarization Phenomena in Nuclear Reactions (University of Wisconsin Press), p. 548 (3) Tanifuji,M. and K. Yazaki, 1968, Prog. Theor. Phys. 40, 1023 (4) Mutchler, G.S., W.B. Broste and J.E. Simmons, 1971, Phys. Rev. C3, 1031

This thesis investigates the properties of an atomic system exhibiting a giant Kerr nonlinearity. The atomic energy level scheme involves four energy levels. A three level A subsystem in the atom exhibits the effect of electromagnetically induced transparency (EIT), reducing the spontaneous emission noise. The fourth level leads to an ac-Stark shift of the ground state, which in turn leads to a giant, noiseless Kerr nonlinearity. Two different environments are explored. First, a system comprising of large number of atoms in an optical cavity is analysed. Detailed aspects of noise reduction in this system are investigated. In particular, strong squeezing in the quadrature in phase with the field driving the cavity mode is found, if the effective coupling of light to the atoms is strong. However, the linewidth of the predicted squeezing is found to be very narrow. This is attributed to a very steep linear susceptibility of the atomic medium. Since the widening of the squeezing window is possible only for weaker effective coupling, in turn reducing the squeezing level, a different environment is proposed. This involves a single four level atom, strongly coupled to the cavity mode. In such a strongly coupled system, the most appropriate approach is found to be that formulated in terms of polaritons – composite excitations of the 'atom-cavity molecule'. Adopting the polariton approach, nonclassical correlations in the field leaving the cavity are investigated. Strong photon antibunching is found and the effect of photon blockade predicted and described. The photon blockade effect can also be found in a system comprised of a two level atom coupled to the cavity mode, if the external driving is tuned to one of the vacuum Rabi resonances. A comparison between the two schemes is performed, and it is found that the four level scheme exhibits much better photon blockade. The reason for this is quantum interference between secondary transitions in the dressed states picture. Destructive interference cancels the transitions that would otherwise introduce a second photon into the system, hence producing a more robust photon blockade. All of these results are valid in the regime where external driving is weak. If the external driving strength is increased, the photon statistics (as measured by the zero-delay second order correlation function) changes from strong antibunching to strong bunching, over a relatively narrow range of driving strengths. The occurrence of this change can again be attributed to quantum interference. It is shown that the interference effect prevents the excitation of the composite system by a second photon, but not excitation by a two-photon transition (following the first excitation). Therefore, the third excitation manifold is excited, which then decays back to the first manifold in a two photon cascade. This two photon cascade is the source of correlated photon pairs causing an increase in the second order correlation function. The dynamics of forward scattering of light is presented, and nonclassical behaviour of the delay dependence of correlation function ('overshoots' and 'undershoots') is discussed. For the analytical treatment of this system, a method based on the polariton approach is devised, which includes the treatment of driving and damping. It is shown that this method is ideally suited to the analysis of strongly coupled systems, where only a few photons contribute to the dynamics.

The proton polarization in the 3He(d,p)4He reaction induced by unpolarized deuterons has been measured at deuteron lab. energies of 2.0, 2.8, 3.9 and 6.0MeV for 20 angles between 0° and 150° (c.m.). Statistical uncertainties are typically ± 0.01. The measurements were made with a proton polarimeter in which the left-right asymmetry of scattering at 60° (lab.) in 4He is determined. The polarimeter employs "venetian-blind" collimation of the protons by conical vanes and 75 cm2 plastic scintillator detectors. Four detectors are included for use in polarization transfer experiments. For 10.5 MeV protons and a helium pressure of 250 p.s.i. the target thickness is 3 MeV and the efficiency per detector per unpolarized proton incident is 10-4. For each polarimeter detector a triple coincidence with a 15 ns resolving time was required with two scintillator transmission detectors preceding the polarimeter. Spectra of random coincidences were accumulated simultaneously and subtracted. Asymmetries resulting from polarimeter-target misalignment and other geometrical effects are discussed. All results quoted are geometric means of pairs of measurements for 180° rotation of the polarimeter and are also arithmetic means of such measurements to left and right of the 3He target. The absolute analyzing power is estimated by computer simulation of trajectories to be -0.638 ± 0.020 for protons entering at 10.3 MeV. The product of polarization and cross section is fitted to an expansion of first-order associated Legendre polynomials using these results and earlier measurements. Only four terms are required except at 6.0MeV where a fifth is necessary. The energy dependence of these coefficients suggests resonances in 5Li at deuteron energies of 60MeV (odd coefficients) and 7.5 MeV (even coefficients) in agreement with results for the polarized-beam analyzing powers(1). Comparison of the results with vector-polarized-beam (1) and polarized-target(2) analyzing powers shows no evidence for the postulated simple relations(3) based on DWBA cal calculations. Comparison of the results with recent measurements of the neutron polarization in the mirror reaction(4) shows no significant differences. The theory of angular correlations in charged particle reactions is developed and used to calculate outgoing nucleon polarizations. Expressions are given for polarization transfer coefficients. These coefficients are evaluated in terms of the T-matrix elements for the interference of various channels with the dominant S-wave, JΠ = 3+/2 channel in 3He(d,p)4He at the 0.43 MeV resonance. Two experiments to measure combinations of these elements are discussed. (1) Gruebler, W. et al., 1971, Nucl. Phys. Al76, 631 (2) Leemann, Ch., W. Gruebler et al., 1971, in Polarization Phenomena in Nuclear Reactions (University of Wisconsin Press), p. 548 (3) Tanifuji,M. and K. Yazaki, 1968, Prog. Theor. Phys. 40, 1023 (4) Mutchler, G.S., W.B. Broste and J.E. Simmons, 1971, Phys. Rev. C3, 1031

This thesis investigates the properties of an atomic system exhibiting a giant Kerr nonlinearity. The atomic energy level scheme involves four energy levels. A three level A subsystem in the atom exhibits the effect of electromagnetically induced transparency (EIT), reducing the spontaneous emission noise. The fourth level leads to an ac-Stark shift of the ground state, which in turn leads to a giant, noiseless Kerr nonlinearity. Two different environments are explored. First, a system comprising of large number of atoms in an optical cavity is analysed. Detailed aspects of noise reduction in this system are investigated. In particular, strong squeezing in the quadrature in phase with the field driving the cavity mode is found, if the effective coupling of light to the atoms is strong. However, the linewidth of the predicted squeezing is found to be very narrow. This is attributed to a very steep linear susceptibility of the atomic medium. Since the widening of the squeezing window is possible only for weaker effective coupling, in turn reducing the squeezing level, a different environment is proposed. This involves a single four level atom, strongly coupled to the cavity mode. In such a strongly coupled system, the most appropriate approach is found to be that formulated in terms of polaritons – composite excitations of the 'atom-cavity molecule'. Adopting the polariton approach, nonclassical correlations in the field leaving the cavity are investigated. Strong photon antibunching is found and the effect of photon blockade predicted and described. The photon blockade effect can also be found in a system comprised of a two level atom coupled to the cavity mode, if the external driving is tuned to one of the vacuum Rabi resonances. A comparison between the two schemes is performed, and it is found that the four level scheme exhibits much better photon blockade. The reason for this is quantum interference between secondary transitions in the dressed states picture. Destructive interference cancels the transitions that would otherwise introduce a second photon into the system, hence producing a more robust photon blockade. All of these results are valid in the regime where external driving is weak. If the external driving strength is increased, the photon statistics (as measured by the zero-delay second order correlation function) changes from strong antibunching to strong bunching, over a relatively narrow range of driving strengths. The occurrence of this change can again be attributed to quantum interference. It is shown that the interference effect prevents the excitation of the composite system by a second photon, but not excitation by a two-photon transition (following the first excitation). Therefore, the third excitation manifold is excited, which then decays back to the first manifold in a two photon cascade. This two photon cascade is the source of correlated photon pairs causing an increase in the second order correlation function. The dynamics of forward scattering of light is presented, and nonclassical behaviour of the delay dependence of correlation function ('overshoots' and 'undershoots') is discussed. For the analytical treatment of this system, a method based on the polariton approach is devised, which includes the treatment of driving and damping. It is shown that this method is ideally suited to the analysis of strongly coupled systems, where only a few photons contribute to the dynamics.