Single-frequency oscillation of thin-disk lasers due to phase-matched pumping

Christian Vorholt; Ulrich Wittrock

doi:10.1364/OE.25.021388

1. Introduction

Single-frequency oscillation of Yb:YAG thin-disk lasers without etalons or other spectral filters is usually limited to low output powers because above a certain pump power additional longitudinal modes reach the lasing threshold and oscillate simultaneously. It seems that the highest reported single-frequency output power of a thin disk laser is 30 W [1]. Single-frequency oscillation and wavelength tuning was achieved by placing a Lyot filter in the laser cavity. In this paper, we present a pump concept for thin-disk lasers that leads to single-frequency emission without etalons or spectral filters. Additionally, the emission wavelength can be tuned in a range of some nanometers by slightly changing the pump arrangement. Since the oscillation wavelength is tunable, the output power of several thin-disk lasers can be combined by a dispersive element to increase the brightness while preserving the spatial beam quality. First experimental results using an intra-cavity pumped thin-disk laser using this pump concept demonstrated single longitudinal mode operation of a thin-disk laser and wavelength tunability of the lasing mode [2].

Thin-disk lasers use an active medium that has the shape of a disk with a typical thickness d of 100 μm to 300 μm and a diameter of several millimeters [3]. The front side of the disk is AR-coated for the pump light and the laser light and the back side of the disk is HR-coated. Thin-disk lasers employ multipass pumping schemes in order to absorb most of the pump light. The most common multipass pumping scheme consists of a parabolic mirror and roof prisms. This setup re-images the pump spot several times onto the disk [4]. Inside the disk, incident pump light and reflected pump light spatially overlap. If the pump light were fully coherent, a standing wave pattern with 100 % contrast would be produced throughout the crystal. In the case of partially coherent pump light, the standing wave pattern has 100 % contrast at the HR-coated back side of the disk and the contrast decreases with increasing distance from the back side. The decrease in contrast depends on the spatial beam quality and the coherence length of the pump light.

Recently, pump diodes became available for pumping the zero phonon line of Yb:YAG at 969 nm [5]. These pump diodes have an emission bandwidth of 0.6 nm. The small emission bandwidth is necessary because the width of the zero phonon line in Yb:YAG is only 0.5 nm [6]. The emission bandwidth of 0.6 nm corresponds roughly to a coherence length of $ℓ_{c} = λ_{p}^{2} / δ λ \approx 1.6 mm$ , where λ_p is the diode wavelength and δλ is the emission bandwidth. This coherence length is roughly twelve times larger than a 130 μm thick disk. As a consequence, the incident pump light and the reflected pump light produce a standing wave pattern throughout the disk with high contrast.

The distance the pump light has to travel between successive double-passes is typically many centimeters and thus much longer than the coherence length. Therefore the intensity that results from the superposition of all double-passes exhibits temporal fluctuations due to the irregular phase relationship between pump light belonging to different double-passes. However, these fluctuations are on the gigahertz scale and therefore irrelevant for the much slower gain dynamics. Note that although the magnitude of the intensity is fluctuating, the period and the position of the standing wave pattern is stationary.

We now show that the standing wave pattern of the pump light can be matched to the standing wave pattern of the laser light. For the moment, we assume that the pump light is a monochromatic plane wave and we neglect absorption. We also assume that the HR coating at the back side of the disk induces the same phase shift on reflection for the laser light and the pump light and that this phase shift is equal for s-polarized light and of p-polarized light. Our assumptions on the phase shifts are usually valid for broadband HR coatings. The pump light hits the disk at an angle θ₀ with respect to the surface normal and propagates at an angle θ inside the disk. The angle θ is related to the angle of incidence θ₀ via Snell’s law θ = sin⁻¹ (sin θ₀/n_p), where n_p is the refractive index of the disk for the pump light. The intensities of the s-polarized standing wave pattern $I_{p}^{(s)}$ and the p-polarized standing wave pattern $I_{p}^{(p)}$ of the pump light are given by [7]

I_{p}^{(s)} (z) = 4 I_{p, 0}^{(s)} {sin}^{2} (k_{p} z cos θ)

I_{p}^{(p)} (z) = 4 I_{p, 0}^{(p)} [{sin}^{2} (k_{p} z cos θ) {cos}^{2} θ + {cos}^{2} (k_{p} z cos θ) {sin}^{2} θ],

where

I_{p, 0}^{(s)}

and

I_{p, 0}^{(p)}

are the intensities of the incident s-polarized and p-polarized pump light of the first pass, k_p is the angular wavenumber of the pump wavelength in the disk given by k_p = 2πn_p/λ_p, and z is the position along the axis of the disk. The grating period of the standing wave pattern is Λ_p = λ_p/(2n_p cos θ). The standing wave pattern of the s-polarized pump light is shown in Fig. 1.

Fig. 1 The left side shows a typical thin-disk that is pumped by multiple pump beams that hit the disk under the same angle of incidence. If the coherence length of the pump light is larger than the thickness of the disk but shorter than the propagation length between two passes, a simple standing wave pattern in the disk arises. The standing wave pattern of the pump light is shown on the right side for an angle of incidence of θ₀. It is assumed that the pump light is fully coherent and s-polarized for the purpose of illustration. After refraction at the AR-coated front side, the pump light propagates at an angle of θ inside the disk. The thickness of the disk is d and the period of the standing wave pattern is Λ_p = λ_P/(2n_p cos θ).

Download Full Size | PDF

We restrict the discussion to I-resonators, i.e. a resonator with a disk and a mirror, creating a straight, unfolded beam such that disk, beam, and mirror together resemble the letter “I”. In an I-resonator the standing wave pattern of the laser light in the disk is given by

I_{L} (z) = 4 I_{L, 0} {sin}^{2} (k_{L} z),

where I_L,0 is the circulating intensity inside the resonator, k_L is the angular wavenumber of the laser light given by k_L = 2πn_L/λ_L, n_L is the refractive index for the laser light, and λ_L is the laser wavelength in vacuum. The standing wave pattern of the laser light has the grating period Λ_L = λ_L/(2n_L). The grating period of the laser light Λ_L and the grating period of the pump light Λ_p are phase-matched when the trigonometric argument k_p z cos θ in Eq. (1) and Eq. (2) is equal to the trigonometric argument k_L z in Eq. (3). For p-polarized light, this condition only leads to optimum phase matching, not to perfect phase matching because the second, weak sin² θ term in Eq. (2) is out of phase with respect to the strong first cos² θ term. Solving for θ gives the phase matching angle inside the disk θ_PM:

θ_{PM} = {cos}^{- 1} (\frac{λ_{p} n_{L}}{λ_{L} n_{p}}) .

According to Snell’s law the angle of incidence of the pump light on the disk θ_0,PM to achieve phase matching is given by

θ_{0, PM} = {sin}^{- 1} [\frac{1}{λ_{L}} \sqrt{λ_{L}^{2} n_{p}^{2} - λ_{P}^{2} n_{L}^{2}}] .

At a pump wavelength of 969 nm and a laser wavelength of 1030 nm, the angle of incidence of the pump light that leads to phase matching is θ_0,PM ≈ 38°. Using this angle of incidence, the medium is not pumped at the nodes of the standing wave pattern of the laser light. These nodes stabilize single-frequency emission if a quasi-three-level medium is used for the disk [8]. Unpumped regions in a quasi-three-level medium represent a loss. A mode whose standing wave is not phase matched to the standing wave of the pump light would have a non-vanishing amplitude of its electrical field at these unpumped regions.

2. Requirements on the pump diodes

Our pump concept requires a standing wave pattern with high contrast inside the disk. The contrast is reduced when the pump light is not polarized, the emission bandwidth of the pump diodes leads to a coherence length that is shorter than the disk’s thickness, and when the spatial beam quality of the pump diodes is poor. We discuss these three influences separately. The Michelson contrast K = (I_max − I_min)/(I_max + I_min) is used as a figure of merit for the contrast.

2.1. Polarization

Conventional thin-disk lasers are pumped by unpolarized light. Hence the diode pump intensity I_p is made up of equal contributions of s-polarized and p-polarized light $I_{p} (z) = I_{p}^{(s)} (z) / 2 + I_{p}^{(p)} (z) / 2$ with $I_{p, 0}^{(s)} = I_{p, 0}^{(p)}$ in the temporal average. According to Eq. (1) and Eq. (2), the intensity at a node of the combined standing wave pattern of the two polarizations is I_min (θ) = 2I_p,0 sin² θ and the largest intensity is I_max (θ) = 4I_p,0. Thus, the Michelson contrast for unpolarized pump light at an angle of incidence θ₀ onto the disk is

K_{unpol} = \frac{2}{{sin}^{2} θ + 1} - 1 = \frac{2}{n_{p}^{2} {sin}^{2} θ_{0} + 1} - 1 .

At an angle of incidence of θ_PM = 38° the contrast of the standing wave pattern reduces to K_unpol(θ_PM) = 0.8 as shown in Fig. 2(a).

Fig. 2 Standing wave pattern of the pump light. In (a) random polarization of the pump light reduces the contrast of the standing wave pattern everywhere in the crystal. (b) An emission bandwidth of the pump diodes that is equal to the critical bandwidth δλ_crit leads to a gradual decrease of the contrast starting from the HR coating at z/d = 0 towards the AR coating z/d = 1. (c) An angular distribution that is equal to the critical angular distribution δθ_crit leads to a similar gradual decrease of the contrast as in (b). The wavelength is exaggerated for the purpose of illustration.

Download Full Size | PDF

2.2. Emission bandwidth

The pump diodes must have a coherence length that is longer than the thickness of the disk which requires a sufficiently small spectral emission bandwidth. We call the emission bandwidth at which the contrast of the standing wave pattern has dropped to zero at the AR-coated side of the disk the critical emission bandwidth δλ_crit. The critical emission bandwidth is calculated for pump light with a top-hat spectral power density. The time averaged standing wave pattern for an emission bandwidth δλ_p of the pump diodes is then approximately given by

I_{δ λ_{p}} (z) = \int_{λ_{p} - δ λ_{p} / 2}^{λ_{p} + δ λ_{p} / 2} \frac{I_{p}}{δ λ_{p}} {sin}^{2} [\frac{2 π n_{p}}{λ^{'}} z cos θ] d λ^{'}

\begin{array}{l} \approx \frac{1}{4} I_{p} {- 2 cos [2 π n_{p} (\frac{1}{λ_{p} + δ λ_{p} / 4} - \frac{1}{λ_{p} - δ λ_{p} / 4}) z cos θ] \\ \times cos [2 π n_{p} (\frac{1}{λ_{p} + δ λ_{p} / 4} + \frac{1}{λ_{p} - δ λ_{p} / 4}) z cos θ] + 2} for δ λ_{p} \leq δ λ_{crit}, λ_{p} > \frac{δ λ_{p}}{2} \end{array}

where I_p is the pump intensity and λ_p is the center wavelength of the pump light. The approximation that was made for deriving Eq. (8) is valid for bandwidths smaller than the critical bandwidth and for a pump wavelength larger than half the emission bandwidth. Eq. (8) is a beat between two spatial frequencies. The beat frequency is approximately given by

f_{δ λ, beat} \approx \frac{δ λ_{P}}{λ_{p}^{2}} n_{p} cos θ

Using this approximation, the envelope functions of Eq. (8) can be written as

I_{δ λ_{p}, + env} (z) = \frac{1}{2} I_{p} [cos (π \frac{δ λ_{p}}{λ_{p}^{2}} n_{p} z cos θ) + 1]

I_{δ λ_{p}, - env} (z) = - \frac{1}{2} I_{p} [cos (π \frac{δ λ_{p}}{λ_{p}^{2}} n_{p} z cos θ) - 1],

where I_{δλ_p,+env} denotes the upper envelope and I_{δλ_p,−env} denotes the lower envelope of the standing wave pattern given by Eq. (8). We can calculate the critical bandwidth δλ_crit by setting Eq. (10) equal to Eq. (11) at z = d. This means the contrast is zero at the AR-coated side of the disk. The critical emission bandwidth is then given by

δ λ_{crit} \approx \frac{λ_{p}^{2}}{2 d n_{p} cos θ} = \frac{λ_{p}^{2}}{2 d {(n_{p}^{2} - {sin}^{2} θ_{0})}^{1 / 2}} for λ_{p} > \frac{δ λ_{crit}}{2} .

The standing wave pattern inside the disk for an emission bandwidth equal to the critical bandwidth is shown in Fig. 2(b). The contrast of the standing wave pattern inside a disk made of Yb:YAG which is 130 μm thick drops to zero inside the disk, if the emission bandwidth of the pump diodes is equal to or larger than δλ_crit = 3 nm for a pump wavelength of 969 nm. If the emission bandwidth is smaller than δλ_crit, the contrast at the AR-coated side of the disk is

K_{δ λ_{p}} (δ λ_{p}) \approx cos (π d n_{p} δ λ_{p} cos θ / λ_{p}^{2}) .

Currently, pump diodes are available having an emission spectrum of even less than 0.6 nm, which is roughly five times smaller than the critical bandwidth of a 130 μm thick disk.

2.3. Spatial beam quality

A full analysis for the required spatial beam quality of the pump diodes is beyond the scope of this paper but a simplified argument can be given by considering plane waves hitting the disk under slightly different angles of incidence. Each of these plane waves leads to a standing wave pattern with a slightly different grating period. In the time average these standing wave patterns can be summed up. The contrast of the resulting standing wave pattern gradually decreases from its maximum value at the HR-coated back side towards the AR-coated front side of the disk. We define the critical width of the angular distribution corresponding to the width of the angular distribution at which the contrast of the standing wave pattern has dropped to zero at the front side of the disk. We assume a monochromatic beam with a homogeneous distribution of power over the angles. The standing wave pattern for pump light having an angular distribution with a width of δθ is given by

I_{δ θ} (z) = \int_{θ - δ θ / 2}^{θ + δ θ / 2} \frac{I_{p}}{δ θ} {sin}^{2} [\frac{2 π n_{p}}{λ_{P}} z cos θ^{'}] d θ^{'}

\begin{array}{l} \approx \frac{1}{4} I_{p} (- 2 cos {\frac{2 π n_{p}}{λ_{P}} [\frac{1}{cos (θ + δ θ / 4)} - \frac{1}{cos (θ - δ θ / 4)}] z {cos}^{2} θ} \\ \times cos [\frac{2 π n_{p}}{λ_{P}} [\frac{1}{cos (θ + δ θ / 4)} + \frac{1}{cos (θ - δ θ / 4)}] z {cos}^{2} θ] + 2) for δ θ \leq δ θ_{crit}, θ > \frac{δ θ}{2} \end{array}

Again, the integral in Eq. (14) is approximately a beat between two close frequencies if the angle θ is larger than half the width of the angular distribution. Then the beat frequency is given by

f_{δ θ, beat} \approx \frac{n_{p}}{λ_{P}} δ θ sin θ

The envelope functions of Eq. (15) are then given by

I_{δ θ, + env} (z) = \frac{1}{2} I_{p} [cos (\frac{π}{λ_{P}} n_{p} z δ θ sin θ) + 1]

I_{δ θ, - env} (z) = - \frac{1}{2} I_{p} [cos (\frac{π}{λ_{P}} n_{p} z δ θ sin θ) - 1]

Then we calculate the critical angular distribution δλ_crit by setting Eq. (17) equal to Eq. (18) at z = d. The critical angular distribution is then given by

δ θ_{crit} \approx \frac{λ_{p}}{2 d n_{p} sin θ} = \frac{λ_{p}}{2 d sin θ_{0}} for θ > \frac{δ θ_{crit}}{2} .

Eq. (15) is depicted in Fig. 2(c) for the critical width of the angular distribution. At a pump wavelength of 969 nm, an angle of incidence of θ₀ = 38°, and disk thickness of d = 130 μm the critical width of the angular distribution is δθ_crit ≈ 0.35°. The critical angle is a function of the angle of incidence because larger angles of incidence lead to a smaller grating period which has the same impact on the standing wave pattern as a larger wavelength difference between laser wavelength and pump wavelength. Therefore phase matching is harder to achieve for large angles of incidence.

The contrast at the AR-coated side of the disk for an angle of incidence θ₀ > 0 and angular distributions having a width smaller than δθ_crit is

K_{δ θ} \approx cos (\frac{π δ θ n_{p} d sin θ}{λ_{p}}) = cos (\frac{π δ θ d sin θ_{0}}{λ_{p}}) for θ > δ θ and δ θ < δ θ_{crit} .

In a thin-disk laser the pump beam is usually focused onto the disk, creating a waist radius w_p at the AR-coated side of the disk that is typically 30 to 50 times the thickness of the disk. Hence, a low beam parameter product is required. The minimum beam propagation factor M² [9] can be obtained that is necessary to achieve the critical angular distribution for a given radius of the pump spot w_p:

M_{min}^{2} \approx \frac{π}{λ_{P}} w_{p} \frac{1}{2} δ θ_{crit} = \frac{π}{4 sin θ_{0}} \frac{w_{p}}{d}

At a pump wavelength of 969 nm, a typical pump spot radius of 3 mm, an angle of incidence of θ₀ = 38°, and disk thickness of d = 130 μm the M² of the pump light should not exceed

M_{min}^{2} \approx 29

. At the moment fiber coupled pump diodes are available with a fiber core of 105 μm in diameter and a numerical aperture NA of 0.22. These pump diodes have an output power around 135 W and a M² = πr_coreNA/λ_P of ≈ 37, where r_core is the radius of the fiber core. The good spatial beam quality allows more pump passes through the disk hence the disk can be thinner and the required spatial beam quality is even lower. These pump diodes could be used in proof-of-concept experiments since they have an emission bandwidth of 0.7 nm in addition to the good spatial beam quality.

3. Analytical examination of output power and optical efficiency

In this section we analyze the stationary rate equations to compare the output power and the efficiency of thin-disk lasers which are pumped homogeneously and those pumped with a standing wave pattern. The laser operates in transverse multimode operation, so that the laser beam has the same diameter as the pump spot on the disk. Additionally, the resonance frequencies of the transverse modes differ only slightly from each other, i.e. we assume that different transverse modes have the same grating period as the fundamental mode. The pump spot is assumed to have a flat top intensity profile. Three different lasers are compared: The first laser is homogeneously pumped and the laser mode can extract the entire gain. This model corresponds to a zero-dimensional, monochromatic rate equation model. The second laser is pumped homogeneously as well but the gain is only extracted by a single longitudinal mode. This case corresponds to a conventionally pumped thin-disk laser that is forced to operate at a single longitudinal mode by using spectral filters. The third laser is pumped by a fully modulated standing wave pattern and the standing wave pattern of the pump is phase-matched to the standing wave pattern of the laser light.

Quantities belonging to the first laser are denoted by noSHB, to the second laser by SHB and to the third laser by PM meaning no spatial hole burning, spatial hole burning, and phase-matched, respectively. All three lasers are pumped with multiple pump passes. The lower case letter i stands for the intensity normalized to the saturation intensity. The saturation intensity at the vacuum wavelength λ is given by I_sat(λ) = hc₀/ [λ (σ_abs(λ) + σ_em(λ)) τ_us], where τ_us is the upper state lifetime, h is Planck’s constant, σ_abs(λ) and σ_em(λ) are the effective cross sections for absorption and emission at the wavelength λ, and c₀ is the speed of light in vacuum.

The mean saturable absorption $\bar{α d}$ at the pump wavelength of the three lasers is given by [10]

{(\bar{α d})}_{noSHB} = \int_{0}^{d} α_{noSHB} (z) d z = α_{p, 0} \frac{(1 - 1 / B) 2 i_{L, 0} + 1}{1 + 2 i_{L, 0} + i_{p, d} Q_{noSHB}} d

{(\bar{α d})}_{SHB} = \int_{0}^{d} α_{SHB} (z) d z = α_{p, 0} \int_{0}^{d} \frac{(1 - 1 / B) 4 i_{L, 0} {sin}^{2} (k_{L} z) + 1}{1 + i_{p, d} Q_{SHB} + 4 i_{L, 0} {sin}^{2} (k_{L} z)} d z

{(\bar{α d})}_{PM} = \int_{0}^{d} α_{PM} (z) d z = α_{p, 0} \int_{0}^{d} \frac{(1 - 1 / B) 4 i_{L, 0} {sin}^{2} (k_{L} z) + 1}{1 + (2 i_{p, d} Q_{PM} + 4 i_{L, 0}) {sin}^{2} (k_{p} z)} d z

where Q_noSHB, Q_SHB and Q_PM are pump enhancement factors (see Eq. (28), Eq. (29) and Eq. (30)). α_p,0 is the small-signal absorption coefficient for the pump light given by α_p,0 = σ_abs,pn_dot at the pump wavelength λ_p, where σ_abs,p is the effective cross section for absorption and n_dot is the doping concentration of the disk. i_L,0 is normalized to the circulating laser intensity inside the resonator and i_p,d is the normalized diode pump intensity, which is the intensity of the light emitted from the pump diodes when it first hits the disk. B is a ratio of the cross sections which we define as follows

B : = \frac{σ_{abs, p} [σ_{abs, L} + σ_{em, L}]}{σ_{abs, L} [σ_{abs, p} + σ_{em, p}]}

where σ_abs,p and σ_em,p are the effective cross sections for absorption and emission at 969 nm and σ_abs,L and σ_em,L are the effective cross sections for absorption and emission at 1030 nm. For our calculations, we use measured effective absorption and emission cross sections at 300 K from Koerner et al. [6]. The integration of Eq. (23) and Eq. (24) is done for d ≫ 1/k_L and d ≫ 1/k_p and yields

{(\bar{α d})}_{SHB} = α_{p, 0} \frac{1 + (1 - B) i_{p, d} Q_{SHB} + {[(1 + i_{p, d} Q_{SHB}) (1 + i_{p, d} Q_{SHB} + 4 i_{L, 0})]}^{1 / 2} (B - 1)}{B {[(1 + i_{p, d} Q_{SHB}) (1 + i_{p, d} Q_{SHB} + 4 i_{L, 0})]}^{1 / 2}} d

{(\bar{α d})}_{PM} = α_{p, 0} \frac{2 i_{L, 0} [1 + (B - 1) {(1 + 4 i_{L, 0} + 2 i_{p, d} Q_{PM})}^{1 / 2}] + B i_{p, d} Q_{PM}}{B (2 i_{L, 0} + i_{p, d} Q_{PM}) {(1 + 4 i_{L, 0} + 2 i_{p, d} Q_{PM})}^{1 / 2}} d .

In the following, the implications of multipass pumping are investigated. For reasons of simplicity, we assume that the pump light is monochromatic and s-polarized for calculating ${(\bar{α d})}_{SHB}$ and ${(\bar{α d})}_{PM}$ . A disk made from Yb:YAG is typically pumped several times above the saturation intensity of the pump light. Pumping leads to a depletion of the ground-state and the absorption of the pump light is reduced. Above the lasing threshold, the ground state is re-populated by stimulated emission of the laser light. We account for the depletion and the repopulation of the ground state by introducing an effective pump intensity. The effective pump intensity is the product of the diode pump intensity i_p,d and the pump enhancement factor Q. For incoherent pump light the spatially averaged effective pump intensity is given by

i_{p, noSHB} = i_{p, d} \underset{: = Q_{noSHB}}{\underset{︸}{\frac{1 - exp (- M_{P} {(\bar{α d})}_{noSHB})}{{(\bar{α d})}_{noSHB}}}},

where M_P is the number of passes of the pump light through the disk. The effective pump intensity for a homogeneously pumped laser in single-frequency operation is given by

i_{p, SHB} = i_{p, d} \underset{: = Q_{SHB}}{\underset{︸}{\frac{1 - exp (- M_{P} {(\bar{α d})}_{SHB})}{{(\bar{α d})}_{SHB}}}} .

In the case of phase-matched pumping the effective pump intensity is a function of the axial coordinate z:

i_{p, PM} = 2 i_{p, d} {sin}^{2} (k z) \underset{: = Q_{PM}}{\underset{︸}{\frac{1 - exp (- M_{P} {(\bar{α d})}_{PM})}{{(\bar{α d})}_{PM}}}} .

These effective intensities are valid for a small single-pass absorption α_p,0d ≪ 1. Note that the enhancement factors Q_noSHB, Q_SHB and Q_PM are a function of the effective pump intensity as well. Thus, the equations are transcendent and a numerical solution of the pump enhancement factors is required.

In the following the mean gain-length products for the three lasers are calculated. The mean gain-length product ${(\bar{g d})}_{noSHB}$ of the homogeneously pumped multimode laser is no function of the axial coordinate and given by

{(\bar{g d})}_{noSHB} = 2 α_{L, 0} \frac{(B - 1) i_{p, noSHB} - 1}{1 + i_{p, noSHB} + 2 i_{L, 0}} d,

where α_L,0 is the small-signal absorption coefficient at the laser wavelength λ_L which is given by α_L,0 = σ_abs,Ln_dot. The mean gain-length product of the laser that is forced to a single frequency

{(\bar{g d})}_{SHB}

and the mean gain-length product of the laser that is pumped by a matched standing standing wave pattern

{(\bar{g d})}_{PM}

is determined by integrating the saturated gain coefficient along the axial coordinate z:

\begin{array}{l} {(\bar{g d})}_{SHB} & = 4 α_{L, 0} \int_{0}^{d} \frac{(B - 1) i_{p, SHB} - 1}{1 + i_{p, SHB} + 4 i_{L, 0} {sin}^{2} (k_{L} z)} {sin}^{2} (k_{L} z) d z \\ \approx \frac{α_{L, 0} d}{i_{L, 0}} [(B - 1) i_{p, SHB} - 1] [1 - {(1 + \frac{4 i_{L, 0}}{1 + i_{p, SHB}})}^{- \frac{1}{2}}], for d k_{L} ≫ 1 \end{array}

\begin{array}{l} {(\bar{g d})}_{PM} & = 4 α_{L, 0} \int_{0}^{d} \frac{(B - 1) i_{p, 0} - 1}{1 + (2 i_{p, d} Q_{PM} + 4 i_{L, 0}) {sin}^{2} (k_{PM} z)} {sin}^{2} (k_{PM} z) d z \\ \approx 2 α_{L, 0} d {\frac{2 i_{L, 0} [{(4 i_{L, 0} + 2 i_{p, d} Q_{PM} + 1)}^{- \frac{1}{2}} + i_{p, d} Q_{PM} (B - 1) - 1]}{{(2 i_{L, 0} + i_{p, d} Q_{PM})}^{2}} \\ + \frac{i_{p, d} Q_{PM} [B {(4 i_{L, 0} + 2 i_{p, d} Q_{PM} + 1)}^{- \frac{1}{2}} - B + i_{p, d} Q_{PM} (B - 1)]}{{(2 i_{L, 0} + i_{p, d} Q_{PM})}^{2}}}, for d k_{p}, d k_{L} ≫ 1 \end{array}

where k_PM = k_P cos θ_PM = k_L is the angular wavenumber in the case of phase-matched pumping. We have not inserted the effective pump power from Eq. (29) into Eq. (32) for reasons of clarity. In steady state operation above the lasing threshold the round-trip gain equals the loss:

exp (\bar{g d}) (1 - T_{OC}) (1 - L_{res}) R_{d} = 1

Depending on the type of laser,

\bar{g d}

in Eq. (34) is replaced by the expressions

{(\bar{g d})}_{noSHB}

,

{(\bar{g d})}_{SHB}

, or

{(\bar{g d})}_{PM}

from Eq. (31), Eq. (32), or Eq. (33), respectively. T_OC is the transmission of the output coupler, L_res denotes the resonator losses due to scattering and impurity absorption, and R_d is the reflectivity of the HR-coated back side of the disk.

The gain-loss balance Eq. (34) is solved for the circulating laser intensity in the resonator i_L. The explicit analytic solution for i_L is complex and not presented here. Fig. 3 shows the output powers and optical efficiencies when Eq. (34) is solved by a fixed-point iteration. For each laser, the output coupling was optimized for a diode pump power of i_p,d = 0.3 and M_p = 16 passes through the disk. At this pump power, the laser that is forced to a single mode has the lowest optical efficiency of 53% in comparison to the other two lasers which have an optical efficiency of 61%. Near the lasing threshold, the homogeneously pumped multimode laser has a similar output power than the homogeneously pumped laser which is forced to a single mode. At the highest investigated diode pump power of i_p,d = 0.3 the homogeneously pumped multimode laser has nearly the same optical efficiency as the single frequency laser that is pumped by the matched pump pattern. The laser single frequency laser that is pumped by the phase-matched standing pump wave has an output power that is roughly 15% higher than the output power of the more incoherently pumped laser that is forced to a single mode. The reason for the lower threshold and the better efficiency of the single-frequency laser that is pumped by the phase-matched standing wave is its better saturation of the gain and its lower reabsorption losses. Less gain is lost as fluorescence because the peaks of the standing wave of the laser mode coincide with the spatial peaks of the gain distribution and there is no reabsorption of laser light at the nodes of the standing pump wave.

Fig. 3 Output intensity normalized to the saturation intensity i_out and optical efficiency i_out/i_p,d of the three lasers. The resonator losses are assumed to be L_res = 0.2% and the reflectivity of the back side of the disk is assumed to be R_d = 1. The output coupling is optimized for a diode pump intensity of i_p,d = 0.3 for each laser individually. The laser that is pumped by a phase-matched standing wave pattern has the lowest lasing threshold and the highest efficiency.

Download Full Size | PDF

In practice, the optical efficiency may be smaller due to energy migration in the crystal that might occur especially in highly doped disks. However, experimental investigations of the spatial hole burning carried out with disks having a doping concentration around 10 % did not suggest a strong influence of energy migration [7].

4. Mode competition in the case of the phase-matched pumping

The mode competition in an Yb:YAG thin-disk laser is simulated based on our previous publication [7]. This simulation iteratively solves the gain-loss balance equation for a given number of modes considering their standing wave patterns. A competition of 300 modes was simulated. These modes were equally spaced in wavelength in the interval from 1027 nm to 1033 nm. A higher number of modes did not change the results.

61 simulations were carried out for different angles of incidence of the pump light. All other quantities were kept constant. The angles of incidence of the pump light ranged from 36° to 40°. The angle of incidence of the pump is coupled to the laser wavelength by Eq. (4). That means the laser oscillates on a mode whose standing wave pattern has the optimal overlap with the standing wave pattern of the pump light. This can be seen in the results of the simulation which are shown in Fig. 4. The black line in the figure indicates the laser wavelength for perfect phase matching. The phase matching condition that is shown by the black line results in laser wavelengths that are significantly shorter than the peak of the spectral gain curve of Yb:YAG if the angle of incidence is larger than 38°. Locking of the laser wavelength by the standing wave of the pump radiation is therefore lost in these regions. The laser trades the advantage of better saturation in the locked state for a better spectral match with the gain. This results in the oscillation of multiple longitudinal modes due to spatial hole burning. The laser oscillates in single-frequency mode in the region from 38.0° to 38.5°. The laser wavelength is locked by the period of the standing wave of the pump and can be tuned by simply changing the angle of incidence of the pump θ₀.

Fig. 4 Simulations of an Yb:YAG thin-disk laser that is pumped by partially coherent pump diodes at 969 nm. The polarization of the pump light is random, the emission bandwidth is equal to δλ = 0.6 nm (FWHM), and the width of the angular distribution is equal to the critical width δθ = δθ_crit. The pump power is 300 W, the pump spot radius is 2.75 mm, the number of double passes is 16, the output coupling is 2 %, and the round-trip intra-cavity loss is 0.4 %. The red crosses indicate the oscillation of a mode at the respective wavelength. The black line depicts the angle, at which the standing wave pattern of a laser mode is equal to the standing wave pattern of the pump mode. (a) shows the resulting optical spectrum of the laser light and (b) shows the output power as a function of the angle of incidence θ₀. The green regions depicts angles of incidence where the laser operates on a single mode only.

Download Full Size | PDF

5. Conclusion

We have presented a new approach for tunable single-frequency emission of thin-disk lasers without using etalons or spectral filters. Sufficiently coherent pump diodes lead to a standing wave pattern inside the disk to which the laser wavelength can be locked. The grating period of this standing wave pattern can be tuned by changing the angle of incidence of the pump light on the disk. This not only leads to a tunable single-frequency laser but also to an increase of the power efficiency of about 10 %. We did not have suitable laser diodes to demonstrate this experimentally. However, we did show tuning of the laser wavelength by phase-matched pumping for an intra-cavity pumped thin-disk laser in a previous publication [2]. Now we have shown that this should also be possible for a diode-pumped thin-disk laser. These results can be used to spectrally combine the beams of several thin-disk lasers that are tuned to slightly different wavelengths.

Funding

Deutsche Forschungsgemeinschaft (DFG) (WI 1939/4-1).

Acknowledgments

We gratefully acknowledge the support of Trumpf GmbH & Co. KG.

References and links

1. C. Stolzenburg, M. Larionov, A. Giesen, and F. Butze, “Power scalable single-Frequency thin disk oscillator,” in Advanced Solid-State Photonics, Technical Digest (Optical Society of America, 2005), paper TuB40.

2. C. Vorholt and U. Wittrock, “Wavelength control by angle-tuning of the laser light in an intra-cavity pumped Yb:YAG thin-disk Laser,” in Advanced Solid State Lasers, OSA Technical Digest (online) (Optical Society of America, 2015), paper AM5A.39. [CrossRef]

3. A. Giesen, H. Huegel, A. Voss, K. Wittig, U. Brauch, and H. Opower, “Scalable concept for diode-pumped high-power solid state lasers,” Appl. Phys. B. 58, 365 (1994). [CrossRef]

4. S. Erhard, A. Giesen, M. Karszewski, T. Rupp, C. Stewen, I. Johannsen, and K. Contag, “Novel pump design of Yb:YAG thin disc laser for operation at room temperature with improved efficiency,” in Advanced Solid State Lasers, OSA Trends in Optics and Photonics, 26, p. 38 (Boston, MA, USA, 1999).

5. B. Weichelt, A. Voss, M. Abdou Ahmed, and T. Graf, “Enhanced performance of thin-disk lasers by pumping into the zero-phonon line,” Opt. Lett. 37(15), 3045–3047 (2012). [CrossRef] [PubMed]

6. J. Körner, C. Vorholt, H. Liebetrau, M. Kahle, D. Kloepfel, R. Seifert, J. Hein, and M. C. Kaluza, “Measurement of temperature-dependent absorption and emission spectra of Yb:YAG, Yb:LuAG, and Yb:CaF₂ between 20 °C and 200 °C and predictions on their influence on laser performance,” J. Opt. Soc. Am. B 29, 2493–2502 (2012). [CrossRef]

7. C. Vorholt and U. Wittrock, “Spatial hole burning in thin-disk lasers,” Appl. Phys. B 120(4), 711–721 (2015). [CrossRef]

8. R. Paschotta, J. Nilsson, L. Reekie, A. C. Trooper, and D. C. Hanna, “Single-frequency ytterbium-doped fiber laser stabilized by spatial hole burning,” Opt. Lett. 22(1), 40–42 (1997). [CrossRef] [PubMed]

9. A. E. Siegman, “Defining, measuring, and optimizing laser beam quality,” Proc. SPIE 1868, 2 (1993). [CrossRef]

10. S. R. Bowman, S. P. O’Connor, S. Biswal, N. J. Condon, and A. Rosenberg, “Minimizing heat generation in solid-state lasers,” IEEE J. Quantum Electron. 46(7), 1076–1085 (2010). [CrossRef]

Single-frequency oscillation of thin-disk lasers due to phase-matched pumping

Abstract

1. Introduction

2. Requirements on the pump diodes

2.1. Polarization

2.2. Emission bandwidth

2.3. Spatial beam quality

3. Analytical examination of output power and optical efficiency

4. Mode competition in the case of the phase-matched pumping

5. Conclusion

Funding

Acknowledgments

References and links

Cited By

Figures (4)

Equations (34)

Optics Express