Cr:Er:Tm:Ho:yttrium aluminum garnet laser exhibiting dual wavelength lasing at 2.1 and 2.9 m: Spectroscopy and laser performance

Transcription

1 JOURNAL OF APPLIED PHYSICS VOLUME 91, NUMBER 1 1 JANUARY 2002 Cr:Er:Tm:Ho:yttrium aluminum garnet laser exhibiting dual wavelength lasing at 2.1 and 2.9 m: Spectroscopy and laser performance Brian M. Walsh a) Department of Physics, Boston College, Chestnut Hill, Massachusetts Keith E. Murray and Norman P. Barnes NASA Langley Research Center, Hampton, Virginia Received 5 July 2001; accepted for publication 24 September 2001 Over 1.0 J of 2.1 m laser energy and over 0.5 J of 2.9 m laser energy have been demonstrated in a single flashlamp pumped solid state laser material, specifically Cr:Er:Tm:Ho:YAG. Flashlamp pumped laser operation of Ho:YAG at 2.1 m and Er:YAG at 2.9 m in various host materials is well known. We have developed an innovative laser system that operates at each of these wavelengths independently or simultaneously in a single solid state laser material with performance comparable to single wavelength systems Er:YAG and Cr:Tm:Ho:YAG. Variation of the flashlamp pump pulse length provides a method to discriminate between lasing at 2.1 and 2.9 m. This effect results from Er Tm Ho energy transfer, the short lifetime of the upper lasing manifold in Er, the 4 I 11/2 manifold, and the relatively long upper laser level lifetime in Ho, the 5 I 7 manifold. This simple tuning method of achieving two widely separated wavelengths without the use of optical tuning elements has potential applications in remote sensing and medical lasers American Institute of Physics. DOI: / I. INTRODUCTION The prospect of producing laser oscillation on two different laser transitions, widely separated in wavelength, in a single laser oscillator provides a motivation to study new codoped laser materials. This article concerns an innovative laser device that has promise for many diverse applications such as medicine and remote sensing. Most lasers operate at one fixed wavelength. To obtain a different, widely separated wavelength often requires changing mirrors to accommodate reflectivity at the desired wavelength and usually changing the laser material to take advantage of an ion with the desired emission properties. For example, Nd:YAG lasing can operate at 946, 1064, and 1330 nm, but one must physically alter the resonator by introducing mirrors which have the reflectivity at the desired wavelength. These Nd transitions also compete with each other since they all share the same upper laser manifold, limiting the efficiency. As a second example, to achieve 2.1 m lasing utilization of Tm:Ho codoped crystals are employed. To achieve 2.9 m lasing Er doped crystals are employed, but this requires a change of gain material. Our device offers a wavelength switchable device simply by changing the pump pulse duration in a Cr:Er: Tm:Ho:YAG crystal. This device is based on the energy transfer dynamics of Er:Tm:Ho and the individual lifetimes of relevant manifolds in Er, Tm, and Ho. Simultaneous dual wavelength lasing has been demonstrated in a number of articles; Er at 2.9 m and Nd at 1.06 m in Er:Nd:YAG, 1 Ho at 3.0 m and Nd at 1.06 m in Ho:Nd:YAG, 2 Nd at 1.06 m and 1.3 m in Nd:YAG, 3 and Nd:YALO 3, 4 Cr at m and Nd at 1.06 m in a Nd:YAG/Cr:LiSAF hybrid. 5 The Er:Nd:YAG and Ho:Nd:YAG two wavelength lasers offer simultaneous lasing on two widely separated wavelengths in two different ions, but suffer from high thresholds and low slope efficiencies, even on the 1.06 m transition due to strong quenching of the Nd 4 F 3/2 upper laser manifold. The Nd:YAG offers simultaneous lasing on two wavelengths in a single ion, and although well separated, the separation is only 250 nm. The Nd:YAG/Cr:LiSAF hybrid offers simultaneous lasing on two ions in two different hosts in the same resonator, but again, although well separated the separation is only 200 nm. This hybrid is certainly an interesting laser, but is not dual wavelength laser operation in a single solid state laser material. The Cr:Er:Tm:Ho:YAG laser we developed offers dual wavelength lasing on two different ions in a single solid state laser material. The laser wavelengths are separated by 1.0 m and low thresholds and good slope efficiencies can be achieved in a simple flashlamp pumped resonator. A unique feature of our Cr:Er:Tm:Ho:YAG dual wavelength laser is the ability to tune the laser simply by changing the flashlamp pump pulse length. By changing the pump pulse length lasing can be achieved individually at 2.1 and 2.9 m or simultaneously at both wavelengths. The advantages of such a device are many. In the field of medicine, hard tissue ablation with 2.9 m laser sources, and soft tissue ablation with 2.1 m laser sources can be used safely. Medical and dental applications are dependent upon tissue absorption for their effectiveness. The absorption of water versus wavelength is shown in Fig. 1. Lasers operating at 2.9 m experience strong absorption in dentin and enamel 6 as well as water, and are thus highly absorbed by teeth and gums. Since the 2.9 m laser energy is strongly absorbed, it interacts with the tissue over very short disa Present address: NASA Langley Research Center, MS 474, Hampton, VA 23681; electronic mail: b.m.walsh@larc.nasa.gov /2002/91(1)/11/7/$ American Institute of Physics

2 12 J. Appl. Phys., Vol. 91, No. 1, 1 January 2002 Walsh, Murray, and Barnes FIG. 1. Logarithmic plot of the absorption depth of water vs wavelength with various laser system operating wavelengths shown. tances. The high absorption allows for the rapid deposition of heat and thus the ability to ablate hard materials. Uses for this wavelength are to ablate thin layers of tissue as in skin resurfacing, or in the dental case, to ablate hard tissues like enamel or dentin, or even old amalgam fillings. Clinical trials have shown that using 2.9 m pulsed lasers are safe when used on pulp and dentin tissue in teeth. They have been shown to be efficient devices for dental caries removal and cavity preparation. Typical energies and repetition rates used in these studies are 80 mj at 5 10 Hz for dental caries removal and 120 mj at 5 10 Hz for cavity preparation. 7 In addition, the use of Er:YAG lasers operating at 2.9 m can be used safely on hard tissues without the use of anesthesia if the laser energy is kept at moderate levels. This is a definite advantage to the patient. Lasers operating around 2.1 m are only moderately absorbed by H 2 O, so it can interact with tissue over longer distances. Lower H 2 O absorption at 2.1 m allows the beam to penetrate and make deeper incisions. Its use as a laser scalpel has been pursued. It also has utility as a dental instrument for gum, pulp, and other soft tissue procedures. Ho:YAG lasers at 2.1 m have been pursued for use in opthamology, urology, orthoscopic surgery, dermatology, and dentistry. Solid-state lasers currently exist for performing medical and dental operatory procedures. There are several disadvantages to the current technologies. The systems currently on the market use two separate lasers to generate the necessary wavelengths. Single wavelength systems are limited in their operatory utility resulting in a high cost for their return. Flashlamp driven, water-cooled systems have relatively good efficiencies and moderately sized overhead. In medical or dental offices, coexisting technologies and limited operatory space make instrument size a major concern. To perform the above described dual mission for soft tissue penetration and ablation of hard materials, two separate laser devices would need to be employed, in essence doubling the space requirements and cost. Our innovation provides both laser wavelengths in a single device with pulse energies for each wavelength that meet the requirements for current medical procedures and surgeries. The medical applications of lasers are so numerous today that a dual wavelength laser device FIG. 2. Illustration of laser wavelength tuning with flashlamp pump pulse length in Er:Tm:Ho:YAG for various time intervals of the pump pulse length. operating with respectable energies at 2.1 and 2.9 m has many potential uses. II. BACKGROUND The initial experiments which led to the discovery of simultaneous lasing on the Er ion around 3.0 m and the Ho ion around 2.0 m in a single solid-state laser material were performed on an Er:Tm:Ho:LuAG flashlamp pumped system. 8 The laser rod contained 10.0% Er, 6.0% Tm, and 0.375% Ho. It was recognized that variation of the pump pulse length allowed either the 2.0 or the 3.0 m to lase individually, or to have them lase simultaneously. In Er: Tm:Ho:LuAG it was found that for flashlamp pump pulse lengths 200 s only Er lased at 2.7 m, while at flashlamp pump pulse lengths 350 s Ho lased at 2.1 m as well. For very long pump pulse lengths, only Ho at 2.1 m lased. A pictorial representation of this tuning with flashlamp pump pulse length appears in Fig. 2. These experiments were performed with double sets of end mirrors. Two mirrors, one coated for high reflectivity at 2.7 m and one for 2.1 m formed the rear mirrors. Two mirrors, one coated for 0.90 reflectivity at 2.7 m and 0.70 reflectivity at 2.1 m formed the front output coupling mirrors. Experiments were conducted to determine the laser performance with flashlamp

3 J. Appl. Phys., Vol. 91, No. 1, 1 January 2002 Walsh, Murray, and Barnes 13 pump pulse length. It was found that optimal slope efficiency could be achieved on the Er 2.7 m laser with a 100 s pump pulse length. These initial experiments produced Er 2.7 m pulses with a 21 J threshold and slope efficiency with a 100 s pulse length, while the Ho 2.1 m pulses produced 72 J threshold and slope efficiency with a 500 s pulse length. The low slope efficiencies are due to the high loss associated with incompatible transmission of the double mirror set at the lasing wavelengths, Er 2.7 m and Ho 2.1 m. In the second phase of the experiments, special dual coated mirrors were used to replace the double mirror sets. 9 These dual band reflective mirrors were designed and fabricated to allow lasing at both 2.7 and 2.1 m. These new mirrors allowed exceptional coalignment of both transitions so that the laser could be tuned simply by changing the pump pulse length. With the dual band mirrors and the Er: Tm:Ho:LuAG flashlamp laser cavity, the Er lased at 2.7 m with a threshold of 29 J and a slope efficiency of with a pump pulse length of 100 s, while the Ho lased at 2.1 m with a threshold of 82 J and a slope efficiency of with a 500 s pump pulse length. This markedly improved performance over the initial demonstration is due to having a single set of dual coated mirrors instead of a double set of singly coated mirrors, which tend to be lossy. This clearly demonstrates the utility of using dual coated mirrors. This experiment proved that dual coated mirrors could be designed to work efficiently on both transitions so that the laser could be tuned to either wavelength simply by changing the pump pulse length. After initial demonstration, secondary experiments served to optimize the Er pump pulse length at 100 s and to demonstrate the use of dual band coated mirrors. However, the performance was modest at best and did not meet the requirements needed for dental procedures. There were two points at issue. One was the lasing wavelength and the other was energy requirements. Addressing the first, a peak in the absorption in dentin and enamel in teeth is at 2.9 m, not 2.7 m. Er:LuAG lases at 2.7 m, but Er:YAG lases at 2.9 m so, it was prudent to change the host from LuAG to YAG. Addressing the second issue of energy requirements required an investigation into the spectroscopy of Er:Tm:Ho codoped systems. This analysis led to the conclusion that high Tm concentrations inhibit the Er lasing at 2.9 m by quenching the lifetime of the 4 I 11/2 upper laser level manifold. Choosing smaller Tm concentrations seemed a prudent choice to increase the efficiency of the Er lasing at 3.0 m. The remainder of this article resumes where these experiments and insights directed us in our research into Er:Tm:Ho two wavelength systems. From these spectroscopic considerations we have reduced the Er lasing threshold by a factor of 4.3 and increased the slope efficiency by a factor of 9.5 while simultaneously improving the Ho laser performance by 20%. III. EXPERIMENT Lifetime measurements to investigate the Er 4 I 11/2 were performed with an alexandrite laser operating at 810 nm for the pump source, corresponding to the Er 4 I 9/2 manifold. The luminescence was detected with a silicon photodiode using a narrow band filter with a center wavelength near 1.0 m. This wavelength corresponds to the Er 4 I 11/2 4 I 15/2 transition and is in a region where neither Tm or Ho exhibit emission. This assures that only the Er luminescence is being monitored, and the effects of the codopants Tm and Ho on the Er 4 I 11/2 lifetime can be unambiguously determined. The decay profiles were monitored with a digital oscilloscope and averaged over 128 traces. Several singly doped Er:YAG samples of different concentrations were examined as well as codoped samples of Er:Ho:YAG, Er:Tm:YAG, and Er: Tm:Ho:YAG. Analysis of the Er 4 I 9/2 lifetime data demonstrates that the presence of Tm and Ho ions shortens the Er decay time, indicating energy transfer mechanisms between Er and the Tm and Ho ions. The laser resonator consisted of a simple straight 22 cm long resonator. For both 2.1 and 2.9 m laser experiments a 2 m concave radius of curvature mirror was used for the high reflector while 0.85, 0.90, and 0.95 reflectivity mirrors were used for the partial reflectors. In a previous publication, 9 special mirrors dual coated to be highly reflective at 2.1 and 2.9 m and a flat output coupler dual coated for reflectivity at 2.1 and 2.9 m were used as proof of concept. These special dual band reflective mirrors were designed and fabricated to allow lasing at both wavelengths in a single resonator. However, the dual coated mirrors were experimental and suffered some damage in the initial experiments. The manufacturer of the mirrors suggested that instead of using fused silica glass as the substrate, that YAG or sapphire might make a better substrate. Since the dual coated mirrors made with YAG or sapphire are considerably more expensive, we used conventional single coating mirrors in the experiments described here, physically changing the mirrors for demonstration of 2.1 or 2.9 m lasing to optimize the Tm concentration and output coupler reflectivity. However, as was shown previously, 9 the dual coated mirrors can be used for exceptional coalignment of both transitions efficiently and to reduce system complexity allowing wavelength tuning by simply changing the pump pulse length. The system was pumped with a single flashlamp in a water cooled, specularly reflecting Kentek cavity. An adjustable pulse length power supply manufactured by Schwartz Electro Optics was used to control the pump pulse length, and hence the lasing wavelength. The cavity cooling temperature was maintained at a nominal 15 C and normal mode laser oscillation was characterized at 1, 2, 5, and 10 Hz. At pulse lengths 200 s lasing on only the Er ion at 2.9 m occurred, while at pump pulse lengths 350 s lasing occurred only on the Ho at 2.1 m. There is some versatility here with regards to the pump pulse length. We used a nominal 100 s pump pulse length for Er lasing at 2.9 m and a 1.0 ms pump pulse length for Ho lasing at 2.1 m. The lasing wavelength dependence on pump pulse length results as a consequence of the short Er 4 I 11/2 lifetime, energy transfer from Er 4 I 13/2 Tm 3 F 4 Ho 5 I 7, and the long lifetime of the Ho 5 I 7 manifold. The lasing wavelength was monitored with a SPEX 0.5 m monochromator coupled to an InAs thermoelectrically cooled photodiode to determine the lasing

4 14 J. Appl. Phys., Vol. 91, No. 1, 1 January 2002 Walsh, Murray, and Barnes TABLE I. Lifetime data of the Er 4 I 11/2 manifold for various sample concentrations. Sample Rise time s Decay time s 30% Er:YAG % Er YAG % Er:YAG % Er:YAG % Er, 1.0% Ho:YAG % Er, 4.0% Tm:YAG % Er, 4.0% Tm, 0.5% Ho:YAG % Er, 5.0% Tm, 1.0% Ho:YAG 2 35 FIG. 3. Energy level diagram showing the energy transfer processes in Er:Tm:Ho:YAG. P ET denotes Er to Tm transfer, P TH denotes Tm to Ho transfer, and denotes nonradiative relaxation. wavelength dependence with pump pulse length. Several 5 85 mm laser rods were characterized. The laser rods were codoped with Cr, Er, Tm, and Ho ions. All the laser rods examined contained 1.0% chromium Cr, 35% erbium Er, and 0.35% Ho. The amount of thulium Tm was different in each rod, with atomic percentages of 1.0%, 0.7%, and 0.4%. IV. SPECTROSCOPY In our attempt to develop a single laser to generate both 2.9 and 2.1 m wavelengths we have investigated the multiply doped system Er:Tm:Ho:YAG. Erbium ions exhibit lasing at 2.9 m onthe 4 I 11/2 4 I 13/2 transition. Ho ions exhibit lasing on the 5 I 7 5 I 8 transition. Tm ions are included in the YAG host because the Tm 3 F 4 manifold provides an efficient energy transfer channel to the Ho 5 I 7. If the flashlamp pump pulse has a short enough duration to deliver a high intensity excitation into the Er 4 I 11/2 manifold, then Er ions will lase at 2.9 m because energy flows into the Er 4 I 11/2 manifold faster than its decay lifetime 100 s. If the pump pulse is longer in duration, then energy flows into the Er 4 I 11/2 manifold more slowly than its natural decay lifetime. In this case, there will be sufficient time for the Er 4 I 11/2 manifold to naturally decay to the Er 4 I 13/2 and act as a donor to the Tm 3 F 4 manifold, which then acts as a donor to the Ho 5 I 7 manifold. The efficient energy transfer process Er 4 I 13/2 Tm 3 F 4 Ho 5 I 7 for long pump pulse duration results in lasing of Ho ions at 2.1 m. Contrary to some of the earlier dual wavelength laser demonstrations, 1 5 the dual wavelengths of Er 2.9 m and Ho 2.1 m laser transitions in Cr:Er: Tm:Ho:YAG operate based on the individual lifetimes of the Er and Ho ions, and their time scale as compared to the time scale of the flashlamp pump pulse length. This is a unique feature of the Cr:Er:Tm:Ho dual wavelength laser. The relevant manifolds and energy transfer processes are shown in Fig. 3 for an Er:Tm:Ho:YAG system. From the above observations, it is clear that lasing on the Er transition, 4 I 11/2 4 I 13/2, at 2.9 m will be affected by the lifetime of the Er 4 I 11/2 manifold, since this represents the storage time for inversion. Energy transfer processes in multiply doped systems are known to be lifetime shortening mechanisms. To investigate the influence of Tm and Ho codopants on the Er 4 I 11/2 lifetime, an alexandrite laser operating at 810 nm was used to excite the Er 4 I 9/2 manifold. Excitations in the Er 4 I 9/2 manifold decay rapidly by nonradiative relaxation to the Er 4 I 11/2 manifold as well as energy transfer processes between Er and Tm ions, most notably Er 4 I 9/2 Tm 3 H 6 Er 4 I 13/2 Tm 3 F 4. This depletes excitations for decay from the Er 4 I 9/2 Er 4 I 11/2, which feeds the Er 4 I 11/2 upper laser manifold. The luminescence decay from the Er 4 I 11/2 4 I 15/2 transition can then be measured around 1.0 m, a region where neither Tm or Ho have emission transitions. The results of these measurements are shown in Table I. From this table it is clear that the Er 4 I 11/2 lifetime is nearly constant with Er concentration in singly doped systems, however the Er 4 I 11/2 lifetime shows some shortening with the addition of Ho and Tm. Since Tm:Ho codoped systems for 2.1 m Ho lasing contain only a small amount of Ho, this is not a large problem, but such systems also contain typically 4% 6% Tm, so the Tm concentration becomes an issue, as seen in Table I. The Er 4 I 11/2 upper laser level lifetime is very short due to nonradiative relaxation since the next lowest lying manifold, the Er 4 I 13/2, is only 3500 cm 1 below the Er 4 I 11/2. The lifetime of the Er 4 I 11/2 is further shortened by the presence of the codopants Tm and Ho. For dopant concentrations around 1.0% Ho the Er 4 I 9/2 lifetime is only decreased by approximately 10%, but for dopant concentrations around 4.0% Tm the Er 4 I 11/2 lifetime is shortened by 50%. There is nothing that can be done to significantly change the nonradiative relaxation decay, however, there is something that can be done to change the Er 4 I 11/2 lifetime in the presence of Tm. By reducing the Tm concentrations to levels below 1.0%, the quenching of the Er 4 I 11/2 upper laser level can be minimized. It must be kept in mind, however, that the primary mechanism for populating the Ho 5 I 7 upper laser level is energy transfer via Tm 3 F 4 Ho 5 I 8 Tm 3 H 6 Ho 5 I 7. For a Tm:Ho only laser, reducing the Tm concentration to such low levels would impede the Ho 5 I 7 lasing because the Tm 3 H 4 absorption would be reduced in addition to seriously inhibiting the 2 for 1 self quenching Tm 3 H 4 Tm 3 H 6 Tm 3 F 4 Tm 3 F 4, which is the main mechanism for populating the Tm 3 F 4 in Tm:Ho lasers. Reducing the Tm concentration in Tm:Ho systems would lead to a reduced population in the Tm 3 F 4 available for direct transfer to the Ho 5 I 7. In the case of an Er:Tm:Ho laser we are not worried about the efficiency of the Tm 3 H 4 self quenching process since the Tm is mainly populated via Er Tm transfer. It is therefore possible to reduce the Tm concentrations to lower

5 J. Appl. Phys., Vol. 91, No. 1, 1 January 2002 Walsh, Murray, and Barnes 15 TABLE II. Comparison of slope efficiency and threshold for Ho 5 I 7 lasing at 2.1 m and Er 4 I 11/2 lasing at 2.9 m. YAG sample m Slope 0.95 R Threshold 0.95 R Slope 0.90 R Threshold 0.90 R Slope 0.85 R Threshold 0.85 R 40.0% Er a J 1.0% Cr, 35.0% Er, 1.0% Tm, 0.35% Ho J J J 1.0% Cr, 35.0% Er, 0.7% Tm, 0.35% Ho J J J 1.0% Cr, 35.0% Er, 0.4% Tm, 0.35% Ho J J J 1.0% Cr, 35.0% Er, 1.0% Tm, 0.35% Ho J J J 1.0% Cr, 35.0% Er, 0.7% Tm, 0.35% Ho J J J 1.0% Cr, 35.0% Er, 0.4% Tm, 0.35% Ho J J J 0.85% Cr, 6.0% Tm, 0.32% Ho b J J a Reference 14. b Reference 13. levels in Er:Tm:Ho lasers without sacrificing Tm 3 F 4 population. In fact, for the Er:Tm:Ho case low Tm concentrations not only minimize Er 4 I 11/2 quenching, but will also increase the fraction of excitations residing in the Ho 5 I 7. The reason is that the fraction of excitations residing in the Ho 5 I 7 increases with a decrease in the ratio of Tm to Ho concentrations, as seen from the equation: 10,11 f Ho C Ho Z 7 /Z 8 exp E Ho ZL C Ho Z 7 /Z 8 exp E Ho ZL C Tm Z 2 /Z 1 exp E Tm ZL, 1 where Z 1,Z 2,Z 7,Z 8 are the partition functions of the Tm 3 H 6, Tm 3 F 4, Ho 5 I 7, and Ho 5 I 8 manifolds, respectively. (kt) 1, and E Ho ZL,E Tm ZL are the so-called zero-line energies of Ho and Tm, respectively. The zero-line energy is the energy difference between the lowest Stark level of the excited manifold and the lowest Stark level of the ground manifold. C Ho and C Tm are the concentrations of Tm and Ho, respectively, in atomic percent. The reader is referred to a previous publication for a derivation. 10 Clearly a prudent concentration balance between Er, Tm, and Ho is needed. Based on the lifetime data, and considerations of the energy transfer dynamics it was decided to have laser rods made with 35% Er, 0.35% Ho, and a range of Tm concentrations at 0.4%, 0.7%, 1.0% atomic. As will be shown in Sec. V, the spectroscopy predictions prove to be quite accurate. By reducing the Tm concentrations from typical values of 4% 6% to less than 1%, the Er 4 I 11/2 laser energy at 2.9 m was increased significantly. The range of Tm concentrations chosen for the laser experiments allowed us to show that optimal performance could be achieved with 0.7% Tm. V. LASER RESULTS Laser performance was evaluated in three laser rods, with dimensions 5 85 mm, each doped with 35% Er, 0.35% Ho, but with three different Tm concentrations, specifically 0.4%, 0.7%, and 1.0%. This range of Tm concentrations were chosen for reasons described in the previous section. For Ho lasing at 2.1 m the pump pulse was set at 1 ms, while for 2.9 m lasing the pump pulse was set at 100 s. For each of the three laser rods with the same Er and Ho concentrations, but different Tm concentrations, a series of output mirror reflectivities were used, 0.95, 0.90, and A set of data was obtained with these parameters for 2.1 and 2.9 m lasing. The slope efficiencies and thresholds for the laser performance are shown in Table II. The thresholds and slope efficiencies of the flashlamp pumped Cr:Er: Tm:Ho:YAG in Table II can be compared to similar flashlamp pumped systems of Cr:Tm:Ho:YAG 12,13 and Er:YAG. Typical Cr:Tm:Ho:YAG flashlamp pumped systems have thresholds of J and slope efficiencies of 0.3% 0.8%. Typical Er:YAG flashlamp pumped systems have thresholds of J and slope efficiencies of 0.1% 1.0%. The variation occurs because of different laser rod dimensions, output coupler reflectivities, and flashlamp pulse lengths. Included in Table II for comparison are results for a mm Er:YAG pumped with a 200 s pulse length 14 and a mm Cr:Tm:Ho:YAG pumped with a 500 s pulse length. 13 As seen in Table II, the dual wavelength Cr- :Er:Tm:Ho laser is compatible in threshold and slope efficiency with both the Er:YAG at 2.9 m and Cr:Tm:Ho:YAG at 2.1 m. Measurements of the lasing performance were also performed as a function of pulse repetition frequency and temperature of the cooling bath. The Ho 2.1 m lasing was highly influenced by temperature. For a temperature decrease of 15 C the laser energy of the Ho 2.1 m increased by 30.0% while the energy of the Er 2.9 m lasing was, for the most part, unaffected. This behavior is due to the fact that Er operates at 2.9 m onthe 4 I 11/2 4 I 13/2 as a four-level laser, while Ho operates at 2.1 m onthe 5 I 7 5 I 8 as quasifourlevel laser. In other words, the lower laser level in the Er 2.9 m laser, an excited state level, is not populated, while the lower laser level in the Ho 2.1 m laser, the ground state level, has a thermal population governed by Boltzmann statistics. In order to achieve a population inversion in the upper laser level it is necessary to overcome any population in the lower laser level. Er 2.9 m lasers terminate on an excited state so there is no lower laser level population to overcome. Ho 2.1 m lasers, on the other hand, terminate on the ground state so there is a significant thermally excited population to overcome depending on the temperature and Boltzmann statistics. The fact that Ho 2.1 m laser energy increases with lower temperature while Er 2.9 m laser energy remains the same is, therefore, reasonable for the reasons just outlined.

6 16 J. Appl. Phys., Vol. 91, No. 1, 1 January 2002 Walsh, Murray, and Barnes FIG. 4. Slope efficiency and threshold for Ho 2.1 m lasing vs temperature of the cooling bath, illustrating the effects of temperature for Ho lasing. Figure 4 shows a plot of the changes in slope efficiency and threshold with temperature for Ho 2.1 m lasing for 1.0 Hz operation with a 1 ms pump pulse length and a 0.95 R output coupler reflectivity. As can be seen in this figure, the slope efficiency increases with lower temperature and the threshold decreases with lower temperature. The slope of a linear fit these curves shows that there is a K 1 increase in slope efficiency and a 1.4 J/K decrease in threshold with decreasing temperature. With respect to pulse repetition frequency, the Ho 2.1 m laser energy decreased by 40% going from 1 to 2 Hz, while the Er 2.9 m energy diminished very little in energy going from 1 to 2 Hz. Er lasing shows much less influence with respect to repetition rate than Ho, decreasing by 20% at 5 Hz and 40% at 10 Hz. This is reasonable since the Er laser transition is driven by a smaller pump pulse length than the Ho transition, 100 s for Er compared to 1000 s for Ho. This order of magnitude change in pump pulse has a profound effect on the thermal lensing characteristics in YAG as well as the population of the lower laser level. A laser rod with 35% Er, 0.35% Ho, 0.7% Tm, and with larger dimensions of mm was also analyzed. For Er lasing there was little change in threshold and slope efficiency as compared to the 5 85 mm rods. For Ho lasing there was a 40% reduction in threshold and a 50% increase in slope efficiency. With a 0.95 R output coupler at 2.1 m, over 1.0 J of 2.1 m laser energy was achieved with 200 J of electrical energy. With a 0.90 output coupler at 2.9 m, over 500 mj of 2.9 m laser energy was achieved with 100 J of electrical energy. The slope efficiencies were and for Er 2.9 m lasing and Ho 2.1 m lasing, respectively. The thresholds were 15 and 50 J for Er 2.9 m lasing and Ho 2.1 m lasing, respectively. The electrical energy in versus laser energy-out is shown in Fig. 5. VI. SUMMARY Over 1.0 J at 2.1 m and over 500 mj at 2.9 m has been achieved in a single solid state Cr:Er:Tm:Ho:YAG laser material. These energies were achieved in normal mode operation at 1 Hz in a mm Cr 1 Er 35 Tm 0.7 Ho 0.35 laser rod. Switching between laser wavelengths can be easily FIG. 5. Comparison of the laser energy vs electrical energy for Ho 2.1 m and Er 2.9 m laser operation in a mm Er 35 Tm 0.7 Ho 0.35 YAG laser rod. achieved by adjusting the pump pulse length. In these experiments it was necessary to change the cavity mirrors also to switch wavelengths, but it has been demonstrated previously that dual coated mirrors can be utilized with comparable results. The Ho lasing results at 2.1 m showed a strong dependence on temperature and pulse repetition rate, while the Er lasing at 2.9 m showed much less dependence on temperature and pulse repetition rate. This behavior can be understood due to the order of magnitude smaller pump pulse length used to initiate Er 2.9 m lasing. Thermal lensing is known to be a problem in YAG, 17 and exhibits itself more in Ho than Er lasing due to the larger pump pulse length and larger pump energies needed to initiate Ho lasing at 2.1 m. A possible method to reduce the deleterious thermal effects on the Ho lasing is to eliminate Cr from laser material. The tradeoffs of increased flashlamp absorption versus thermal heating due to Cr were not tested since funding allocation precluded another growth run for laser rods without Cr. So, a question remains with regards to the utility of adding Cr and its impact on laser performance, especially with regards to Ho lasing at higher repetition rates where heating becomes a major issue. Spectroscopic analysis of the Er 4 I 11/2 lifetime in the presence of codopants Tm and Ho indicated that the presence of these codopants, especially Tm, led to significant lifetime shortening of the Er 4 I 11/2 upper laser level. This was a very important finding that led to an order of magnitude increase in the efficiency of the Er 2.9 m lasing. Several Cr:Er:Tm:Ho:YAG laser rods, based on these findings, were manufactured with smaller than usual Tm concentrations, specifically 0.4%, 0.7%, 1.0% atomic. Using this set of materials and a series of output coupler reflectivities, specifically 0.85, 0.90, and 0.95, we were able to optimize the laser system. Optimal performance for the Er 2.9 m laser prefers 0.4% Tm with a 0.90 R output coupler, while Ho 2.1 m lasing prefers 0.7% Tm and a 0.95 R output coupler. The current energies obtained at 2.1 and 2.9 m from 1 to 5 Hz are certainly within the range of energies and repetition rates currently being used in clinical trials for dental procedures. A 10 Hz repetition frequency is also viable for Er

7 J. Appl. Phys., Vol. 91, No. 1, 1 January 2002 Walsh, Murray, and Barnes m lasing, but is unlikely for 2.1 m lasing without some modification of the resonator to correct for thermal lensing. This work demonstrates a step forward in producing an efficient and cost effective medical laser device that can do the work of two conventional medical laser devices, one at 2.9 m and one at 2.1 m. The energies and performance achieved in Cr:Er:Tm:Ho:YAG at both 2.1 and 2.9 m are comparable to the single wavelength devices Er:YAG at 2.9 m and Cr:Tm:Ho:YAG at 2.1 m. Dual wavelength lasing in Cr:Er:Tm:Ho:YAG has been shown to be an efficient and simple method of achieving two widely separated laser wavelengths in a single solid state laser material. This dual wavelength laser is based on lifetimes of the lasing ions, energy transfer dynamics, and variation of flashlamp pump pulse length. It is hoped that this research opens up ideas and possibilities for new dual wavelength lasers operating on widely separated wavelengths in rare earth codoped systems. ACKNOWLEDGMENT One of the authors would like to acknowledge the sponsorship of this work by NASA Grant Nos. NAG and NAG W. Q. Shi, R. Kurtz, J. Machan, M. Bass, and M. Birnbaum, Appl. Phys. Lett. 51, J. Machan, R. Kurtz, M. Bass, and M. Birnbaum, Appl. Phys. Lett. 51, R. A. Morgan, F. A. Hopf, and N. Peyghambarian, Opt. Eng. 26, W. Lin and H. Shen, J. Appl. Phys. 86, P. P. Lin, M. A. Andriasyan, B. A. Swartz, N. Witherspoon, and J. H. Holloway, Appl. Opt. 38, P. Rechmann, D. S. Goldin, and T. Henning, Lasers Den. IV, SPIE Prog. Biomed. Opt. 3248, C. D. Cozean and L. Powell, Lasers Dent. IV, SPIE Prog. Biomed. Opt. 3248, N. P. Barnes, K. E. Murray, B. M. Walsh, and R. L. Hutcheson, OSA Trends Opt. Photonics Ser. 10, K. E. Murray, N. P. Barnes, B. M. Walsh, R. L. Hutcheson, and M. R. Kokta, OSA Trends Opt. Photonics Ser. 34, B. M. Walsh, N. P. Barnes, and B. Di Bartolo, J. Lumin. 75, B. M. Walsh, N. P. Barnes, and B. Di Bartolo, J. Lumin. 90, S. R. Bowman, M. J. Winings, R. C. Y. Auyeung, J. E. Tucker, S. S. Searles, and B. J. Feldman, IEEE J. Quantum Electron. 27, M. G. Jani, N. P. Barnes, and K. E. Murray, Appl. Opt. 36, S. Wuthrich, W. Luthy, and H. P. Weber, J. Appl. Phys. 68, S. Schnell, V. G. Ostroumov, J. Breguet, and W. A. R. Luthy, IEEE J. Quantum Electron. 26, J. Breguet, A. F. Umyskov, S. G. Semenko, W. Luthy, H. P. Weber, and I. A. Shcherbakov, IEEE J. Quantum Electron. 28, W. Koechner, Solid State Laser Engineering Springer, New York, 1995.