A narrow-band tunable diode laser system with grating feedback
Contents
1. and directed into a photodiode A preliminary assessment of mechanical electrical and thermal stability may be made merely by observing the single beam absorp tion line The photodiode output is converted to a voltage by an I V amplifier whose circuit is shown in Fig 11 and the resulting signal is displayed on an oscilloscope Make sure the I V offset is not set to an extreme value that saturates the amplifier at either the positive or negative supply voltage Next the piezoelectric element should be driven by a triangle wave from an ordinary function gen erator at 15 to 30 Hz with peak to peak amplitude up to 30 V The photodiode signal should vary by 5 50 de pending on the particular cell as the PZT scans the laser across the absorption line It is helpful to trigger the scope from the function generator sync pulse or TTL output or operate the scope in X Y mode in order to obtain a stable display as the PZT drive is adjusted When electrical tun ing of the laser over the absorption line has been obtained it is a good time to reexplore mechanical adjustments of the grating angle and diode drive currents An absorption line or its neighbors corresponding to different hyperfine levels of the ground state or different isotopes recurs sev eral times for nearby currents or grating angles Also dis continuous steps of photodiode output occur across the oscilloscope trace These steps correspond to transitions from one longitudinal e2. A narrow band tunable diode laser system with grating feedback and a saturated absorption spectrometer for Cs and Rb K B MacAdam A Steinbach and C Wieman Joint Institute for Laboratory Astrophysics and the Department of Physics University of Colorado Boulder Colorado 80309 0440 Received 6 February 1992 accepted 26 June 1992 Detailed instructions for the construction and operation of a diode laser system with optical feedback are presented This system uses feedback from a diffraction grating to provide a narrow band continuously tuneable source of light at red or near IR wavelengths These instructions include machine drawings for the parts to be constructed electronic circuit diagrams and prices and vendors of the items to be purchased It is also explained how to align the system and how to use it to observe saturated absorption spectra of atomic cesium or rubidium I INTRODUCTION Tuneable diode lasers are widely used in atomic physics This is primarily because they are reliable sources of narrow band lt 1 MHz light and are vastly less expen sive than dye or Ti sapphire lasers However the frequency tuning characteristics of the light from an off the shelf laser diode is far from ideal and this greatly limits its utility In particular the laser output is typically some tens of MHz wide and can be continuously tuned only over certain limited regions These characteristics can be greatly improved by th
3. homemade components Before discussing the construction aspects we will provide some information concerning the purchas ing of the commercial components We purchase the diode laser itself the collimating lens the fine adjustment screw which controls the lens focus the diffraction grating the mirror mount which holds the grating and the piezoelec tric disks The purchase of most of these items is straight forward Fine adjustment screws and mirror mounts are available as standard items from most companies that sell optics hardware Similarly laser diode collimating lenses and diffraction gratings are available from numerous com panies For the convenience of the reader we list in the Appendix the exact products we use along with the prices and vendors However for these items our choice of ven dors was primarily determined by expediency and we have no reason to think that other vendors would not provide equal or superior products In contrast in order to obtain satisfactory laser diodes and piezo disks we have tried and rejected a large number of different vendors Piezo disks are widely sold as elec tronic speakers and are very inexpensive but most models are not adequate for this application The Appendix gives the only suitable product we have found The purchasing of diode lasers can be filled with frustrations and pitfalls and we refer the reader to Ref 1 for a full discussion of the subject Here we shall just give a b
4. plate As with any laser changes in the length of the cavity cause shifts in the laser frequency Therefore to obtain a stable output frequency undesired changes in the length due to mechanical movement or thermal expansion must be avoided To reduce movements due to vibration of the cavity we mount it on small soft rubber cushions To avoid thermal changes the baseplate is temperature controlled using heaters and or thermoelectric coolers In addition to controlling the temperature of the baseplate we indepen dently control the temperature of the laser diode Finally to avoid air currents interfering with the temperature con trol we enclose the entire laser system in a small insulated metal box Of course to finely tune the laser frequency one must have some way to change the length of the cavity in a carefully controlled manner We do this using a piezo electric transducer speaker disk which moves the grating in response to an applied voltage The laser system also requires a small amount of elec tronics A stable low noise current source is needed to run the laser and temperature control circuits are used to sta bilize the diode and baseplate temperatures This electron ics is readily available commercially However for those 1099 Am J Phys Vol 60 No 12 December 1992 with more time than money we provide circuit diagrams for the relatively simple circuits that we normally use This system contains both purchased and
5. 250 OPEN TO 0 150 x 0 275 ee ee END MILL 0 040 TOP 40 3751 RELIEVE TO 0 020 VIEW 0 188 yy PEIA FOR STABILITY R ki f 0 438 30 BEVEL FOR BEAM CLEARANCE ALIGNMENT JIG 2 64 lt 1 250 A m TAP g Jo 835 0 500 at E F Jo 250 Y gt k 0 125 lt 0 750 gt lt 1 500 0 125 gt Le gt Fig 3 Collimating lens mount and alignment jig machine drawing Di mensions are in inches bridge design has been found to make a significant im provement on laser cavity stability The 0 356 in diam hole to receive the diode package may be made either by boring on a lathe fitted with a four jaw chuck or more easily by a suitable end mill Reground 3 8 in end mills can often be found near this diameter Some deburring or filing may be necessary to allow the diode to fit snugly into its recess but allow it to be rotated to its proper orientation in the initial step of alignment A small hole whose diam eter is selected to fit the thermistor should be drilled into the back side of the mounting block near the diode recess Fig 4 Jig usage in collimation LD laser mounting block Fig 2 C collimating lens mount Fig 3 J alignment jig Fig 3 SP spring or rubber pad to provide a restoring force against adjusting screw S2 K MacAdam A Steinbach and C Wieman 1100 DIODE LASER BASEPLATE SAW SLOT TO FORM HINGE WIDEN TO ALLOW CLEARANC
6. and capped by a lid or a single piece of hollow rectangular tubing may be selected to form the walls Wall thickness should be 1 4 to 1 2 in Inside dimensions about 3 5 in wide 5 5 in long and 4 in high are adequate The floor of the enclosure should stand on some firm support to bring the laser output beam to a desired height above the table The lid of the enclosure should be easily removable to allow frequent access to the laser with minimal disruption of the thermal or mechanical stability After the laser is assembled and satisfactorily aligned drill holes in the box to allow access to the grating adjustment screws and drill an opening to allow exit of the laser beam These steps should be delayed until one knows for certain where the holes should be placed The output aperture is ultimately covered by a microscope slide and the access holes should be plugged to limit air currents Tapped holes on opposite edges of the bottom plate of the enclosure allow the laser structure to be anchored onto the vibration isolation pads discussed below for instance by stretching a rubber band over the laser baseplate and looping it over screw heads in the edges of bottom plate The bottom edge of one of the side walls of the enclosure should be provided with a notch or channel at both ends of the laser for egress of all wires Soft rubber placed in the notches can serve to press the wires firmly against the bot tom plate and in this way the movem
7. and fuse windows onto the ends The optical quality of the windows is unimportant The glass cell is connected to a vacuum system through a glass tube about 1 4 in diameter so that the cell can be evacuated to between 1077 and 10 Torr After the cell is filled it will be tipped off by heating this connecting tube until it collapses in on itself A few grams of alkali metal in a glass ampule are placed in a separate arm on the vacuum system The sys tem should be pumped down and the cell outgassed briefly by heating it for several minutes with a torch At the point where it will be tipped off the glass connecting arm should be repeatedly heated until it just starts to soften but does not collapse in After the outgassing is completed the am pule should be broken to release the alkali metal into the 1107 Am J Phys Vol 60 No 12 December 1992 12K OWA 30K oO W 4 GAIN 100K PD IN Si 300K eeo oN 150 3M omt 100K OFFSET 1M PD OUT 190k S OO es 2 5 ji 1200 1200pF 3 eE F3S6 ee 100K LF356 power pins are bypassed 150 with luF capacitors Fig 11 I to V amplifier circuit system Often when one purchases ampules of alkali metal they come packed with an inert gas In this case there will be a burst of gas also released into the system which must be pumped away The alkali metal can be moved into the cell simply by heating the glass around it and thus distilling it down th
8. for high resolution spectros copy Opt Commun 85 355 359 1991 The user should be aware in evaluating gratings for use that the effi ciency is highly sensitive to polarization See E G Loewen M Nevi re and D Maystre Grating efficiency theory as it applies to blazed and holographic gratings Appl Opt 16 2711 2721 1977 Steve Chu Stanford Univ private communication Steve Swartz Univ of Colorado private communication For a discussion of saturated absorption spectroscopy see W Demtroder Laser Spectroscopy Springer Verlag New York 1981 1992 American Association of Physics Teachers 1111
9. from the lens may weakly filter through the card Probing from all four directions into the collimated beam will indicate both the horizontal and vertical misalignment of the return beam and the objective is to adjust the screws of the grating mount so that the width of the illuminated region on the card edge is brought exactly to zero for each direction of approach A precise vertical alignment of the return beam is made by reducing diode current to just above threshold Then observe the intensity of the output beam while adjusting the tilt of the grating around a horizontal axis If the pre liminary Littrow alignment was adequate the output beam should significantly brighten at the exact vertical position that optimizes feedback into the diode After completing this adjustment the threshold current will be lower than the value recorded earlier for the diode The laser should now be operating with grating controlled feedback near its free running wavelength If more than one vertical setting of the grating appears to enhance the laser output near threshold or if the output beam projected on a distant surface consists of more than a single collimated spot the fault may lie with imperfec tions chips scratches dirt on the grating laser window or lens surfaces VII TUNING THE LASER FREQUENCY A low resolution lt 1 nm grating spectrometer is useful to assess the tuning characteristics of the laser discussed below After initial al
10. gratings at the 852 nm wavelength for Cs can be improved by 10 20 by evaporating a gold coat ing onto the grating before it is installed K MacAdan A Steinbach and C Wieman 1101 GRATING GRATING MOUNT QIZZIZIZITIZZIZITA BDGSSSISSNAA SSS SS CSSSSSSSSS SSS Z227 TIT LTD TO TRIANGLE WAVE GENERATOR SCREW SOFT SOLDER a 4 places ELEMENT Ag plated Fig 6 Piezoelectric disks not to scale Before the grating is attached to the mount the mount should be modified if necessary so that it has the same bridge profile on its base as described earlier In addi tion the grating mount should be modified so that its ad justment screws can be turned by a ball end wrench through holes in the temperature control enclosure of the assembled laser A good way to do this is to remove the heads from 1 4 in 20 socket screws in a lathe and to at tach them to the centers of the knobs of the adjustment screws with epoxy cement E Piezoelectric disks Piezoelectric PZT disks are inserted between the grat ing mount adjustment screw and the movable face of the mount in order to rotate the grating about a vertical axis and alter the cavity length with electrical control Fig 6 Each PZT element consists of a thin brass backing about 1 in in diameter to which a thin smaller diameter silver plated piezoelectric slice is attached in the center with ad hesive around its edge When vol
11. is used to mount the alignment jig The most obvious feature of the baseplate is its flex hinge de sign which allows smooth variation of the spacing between diode and collimation lens by action of a commercial pre cision screw mounted to push against the hinge The slot that forms the hinge can be cut by a bandsaw after all holes are laid out The hole intended to receive the precision adjusting screw should be reamed to allow a fit without excess clearance After all machining of the baseplate is complete the screw can be mounted in this hole with ad hesive or by a set screw The web that provides the flexible hinge should be left 1 16 in or more in width One can always remove material later if it proves too stiff One or more holes should be drilled in the baseplate to mount the diffraction grating holder but the exact position s de pends on the dimensions of the holder and grating and on the desired cavity length Several suitable holes drilled at 1101 Am J Phys Vol 60 No 12 December 1992 this time will allow comparison of laser performance with different cavity lengths without complete disassembly of the collimated laser D Grating and grating mount The baseplate design is intended for use with a 1200 line per mm grating Suitable gratings are readily obtained with 500 and 750 nm blazes and dimensions 1X 1X 3 8 in thick When mounted the grating has its rulings verti cal and diffracts its first order interference ma
12. 15U The resistors on Sl are each 4 99k 1 9 metal film resistors R1 R2 MON SET 1 5K A 38k LN POINT 3 1 THERMISTOR ed LM399 SOK UL ROOM TEMP y e R3 R4 B SOK SOK vA 7 Cl and C2 are 5 C2 l t 17SM iuf Polystyrene capacitors 3 C1 RG TLO 4 luf 15M E THERM MON Ug 13 14 3 12 F TLO74 10U J s O HEAT H xRes stors are a SW G 1 metal f lm 2 Ohm lt 7 Fig 7 Temperature control circuit ing block The temperature control circuit which drives the heater is shown in Fig 7 Although this circuit is rather crude compared to what is usually used for precision tem perature control we have found it adequate for most pur poses It is simply a bridge an amplifier and an RC filter which rolls off the gain as 1 frequency for frequencies between 0 005 and 0 50 Hz The components have been chosen so that above 0 5 Hz the electrical gain is constant This frequency response was selected so that the combina tion of this electrical response and the thermal response of the laser mounting block results in a net servo gain which goes nearly as 1 f The gain is set by the 5 kQ potentiom eter to be just below the point where the servo loop oscil lates One can readily observe saturated absorption spectra and carry out other atomic spectroscopy experiments with temperature control only on the laser mounting block However temperature stabilizing the baseplate greatly re duces the thermal drift of
13. 92728 8020 714 963 9811 13 Thermoelectric cooler 30X30 mm CP1 4 71 O45L 19 00 MELCOR 990 Spruce St Trenton NJ 08648 609 393 4178 14 Cesium and rubidium vapor cells We have never used these cells but this company has announced that they will sell low cost vapor cells to educational institutions Environmental Optical Sensors Inc 3704 N 26th St Boulder CO 80302 303 440 7786 1111 Am J Phys 60 12 December 1992 JILA Visiting Fellow 1991 1992 Permanent address Department of Physics and Astronomy University of Kentucky Lexington KY 40506 0055 IC Wieman and L Hollberg Using diode lasers for atomic physics Rev Sci Inst 62 1 20 1991 J C Camparo The diode laser in atomic physics Phys 26 443 477 1985 R Ludeke and E P Harris Tunable GaAs laser in an external dis persive cavity Appl Phys Lett 20 499 500 1972 M W Fleming and A Mooradian Spectral characteristics of external cavity controlled semiconductor lasers IEEE J Quantum Electron QE 17 44 59 1981 Diode lasers that are supplied without an output window or diodes whose hermetic package has been carefully opened may be AR coated with SiO by the user who has suitable optical coating apparatus and improved operation may result See M G Boshier D Berkeland E A Hinds and V Sandoghdar External cavity frequency stabilization of visible and infrared semiconductor lasers
14. E FOR TIP 2 438 OF DRILL AND REAMER 0 813 D a 0 938 ay N F 4 40 CLR 0 250 0 500 a 3 1 2 4 40 TAP 1 4 DIA REAM HOLE TO ACCEPT PRECISION SCREW BUSHING P DRILL CLEARANCE __ HOLE S FOR SCREW k 1 5 325 TO ATTACH GRATING k 2 5 16 gt MOUNT MATERIALS ALUM 1 2 THK DIMENSIONS IN INCHES Fig 5 Laser baseplate machine drawing B Collimator lens mount Figure 3 shows the aluminum block that holds the flanged collimator lens The placement of the lens axis at 0 500 in above the base and the diameter of the hole are again the critical dimensions The 30 bevel shown on the front side of the block allows clear passage of the output beam off the diffraction grating when very short cavities are used The figure also shows dimensions of a suggested alignment jig that is used to allow transverse displacement of the lens holder without rotation or longitudinal move ment A 2 56 screw with rounded tip and a small piece of bent spring steel or resilient cushion should be prepared for use with the jig as shown in Fig 4 C Baseplate The baseplate is shown in Fig 5 We have found that aluminum is adequate for most purposes and is easy to machine If greater thermal stability is required however the baseplate can be made of invar The two pairs of 4 40 holes should be carefully positioned to match correspond ing holes in the laser and collimator lens blocks The single 4 40 tapped hole
15. II ENCLOSURE AND VIBRATION ISOLATION After all preceding steps of alignment have been com pleted the laser should be thermally and vibrationally iso lated in its enclosure We have achieved an adequate de gree of mechanical isolation by supporting the laser inside the enclosure on three rubber pads that form a tripod un der the solid part of the baseplate avoiding the hinge Soft sorbothane rubber 1 8 in thick cut into 1 2 in squares and stacked to a height 3 8 in forms a springy but well damped support that isolates from vibrations above about 100 Hz For extra isolation additional rubber may be placed under the support which holds the enclosure at the desired height The wires from heaters thermistors the diode and the PZT should be taped down to the laser and or the enclosure baseplates to decouple them mechan ically from the laser cavity The suggested enclosure design offers additional decoupling by pressing the wires firmly against the bottom plate where they exit from the box K MacAdam A Steinbach and C Wieman 1106 DIODE LASER Rb VAPOR CELL PHOTODIODES gt gt Ito V SATURATING AMPLIFIER BEAM Fig 10 Beam layout for saturated absorption Holes should now be drilled in the sidewalls of the box to allow the beam to exit and to allow manual grating adjustments without having to open the enclosure Because of the very nonrigid support of the laser mechanical ad justments alth
16. R CONSTRUCTION OF LASER COMPONENTS Construction of the diode laser system begins in the ma chine shop and primarily requires a milling machine and drill press Detailed machine drawings for the laser mount ing block baseplate collimating lens holder and an align ment jig are given in Figs 2 and 3 In addition an enclo sure should be fabricated but its design is not critical We provide dimensions for mounting the standard Sharp laser package Small changes may be needed for lasers from other vendors In view of the setup time required in ma chining and the fact that many interesting experiments with diode lasers require more than one of them it will probably be found economical to make two or more sys tems at once A Laser mounting block The laser diode is held firmly in a small aluminum block whose details are shown in Fig 2 The critical dimensions are the 0 500 in height of the laser center above the base plate and the depths of holes that ensure that the 9 mm flange of the diode package is gripped by the mounting screws For stability when the block is screwed down to the baseplate the bottom surface of the block should be ma chined as shown with a 0 020 in relief cut down the mid dle so that contact is along the edges of the block This 1100 Am J Phys Vol 60 No 12 December 1992 COLLIMATOR LENS MOUNT 1 000 0 875 0 563 FRONT gt 0 250 SIDE VIEW 0 356DIA TO FIT LENS k 1 125 0
17. arp lasers is at about 839 nm and we have used such lasers to reach the cesium line by heating them However it can be difficult to obtain 839 nm lasers and to obtain reliable performance when tuning the laser this far from its free running wavelength The heating of the laser also degrades its lifetime The remaining components of the laser system are homemade The key components are the laser mounting block which holds the actual diode laser the holder for the collimating lens and the baseplate onto which all the com K MacAdam A Steinbach and C Wieman 1099 LASER DIODE MOUNTING BLOCK ND ace A25 ees ETO a RECEIVE m 0 807 gt DEPTH __ 0 164 THERMISTER 0 563 gt eon SE gt 0 318 j N i SIDE VIEW Lot 4 4y 0 356DIA 0 302D1A 0 235D1A n BOTTOM VIEW ay a rae aes 0 040 ASSEMBLY aa E Re 4 40 SCREEN WASHER CUT AWAY Fo Fig 2 Laser mounting block machine drawing Dimensions are in inches The hole sizes spacings and depths are correct for a Sharp LTO25MDO laser and may be modified for other types ponents are fastened In addition we also make the box that encloses the system and a small jig that is useful for setting the position of the collimating lens All these com ponents have been designed so that they can be constructed by a novice machinist II INSTRUCTIONS FO
18. btained the lens mount screws should be firmly tightened and the jig removed After the lens mount is tightened in place the fine adjustment screw should again be adjusted to precisely collimate the beam Positioning the lens without the jig is also possible for those users with a steady hand but it is very difficult to avoid random rota tions and displacement along the beam when only a trans verse adjustment is desired C Power output and threshold current measurements After the laser has been aligned and collimated and be fore the grating is installed the power output and thresh old characteristics should be recorded and compared with the specifications The threshold current depends on laser temperature so it may be desirable to stabilize the temper ature of the diode mount at this time The output power can be measured as a function of drive current by illumi nating the face of a wide aperture photodiode D Mounting and adjusting the diffraction grating The mounting of the diffraction grating has been de scribed earlier The laser cavity length is determined by the distance from the back of the laser chip to the illuminated spot on the grating and can be as short as about 20 mm in this design The grating mount should be screwed to the laser baseplate to form a cavity of the desired length so that the collimated beam illuminates the center of the grating at K MacAdam A Steinbach and C Wieman 1105 approximately th
19. d laser beam can be easily observed after it has gone 1 5 m or more from the laser The current supply is then set to a normal operating current The output beam at 780 nm when projected onto white paper attached to the wall will hardly be visible with the naked eye but will show up readily in an IR viewer The 852 nm light can only be observed by the viewer or an IR sensitive card At this time interference rings or fringes may be apparent in the projected beam These are normally caused by dust finger prints etc on the laser s output window The window should be cleaned with an optical tissue dampened in methanol so that the beam pattern is uniform and clear The orientation of the diode in its recess should be set either by noting the major axis direction or by checking the output polarization Once the proper orientation is achieved the mounting screws that hold the laser in its recess should be tightened Make sure that connections to the diode including the network of protection diodes are insulated and arranged so that short circuits will not occur during routine handling Note carefully the center position of the dispersed beam spot both its height and lateral po sition and mark it on the wall Despite the broad and un differentiated beam spot the center can be judged reliably within 2 B Collimation The next step is collimation of the output beam and shimming if necessary of the laser or lens mounting blocks to
20. e Littrow angle Best results have been obtained with the shortest possible cavities apparently be cause the corresponding mode spacing about 8 GHz avoids excitation of adjacent cavity modes by the inherent relaxation noise of the diode at around 3 GHz from line center It is possible with a carefully aligned 780 nm laser having 20 mm cavity length to tune electrically over 7 GHz without a mode hop using only the PZT The following procedure is used to align the diffraction grating A small card cut from stiff white paper or a file folder about 2 X 1 4 in is useful as a probe to see that the beam diffracted from the grating returns approximately to the center of the lens The beam spot at 780 nm is readily visible to the eye on the card but at 852 nm the IR viewer is required Before screwing the grating mount firmly to the baseplate in this coarse alignment make sure that the adjustment screws are in midrange The card should be used next to make a more careful alignment of the grating If the return beam is for example too high as the card is lowered vertically in front of the lens the outward face of the card will be illuminated along a narrow region at its edge until the beam is completely cut off The width of this narrow region indicates the degree of vertical misalign ment When an edge of the card is raised from below to cut off the beam no such region of direct illumination will be visible in this example although direct light
21. e glass tubing into the cell It is only necessary to have a few very small droplets in the cell so one ampule is sufficient to fill many cells It is desirable to put much less than 1 g of metal into the cell to reduce the tendency of the metal to coat the windows Once the alkali is in the cell the sidearm is tipped off and the cell is ready for use B Optical setup Figure 10 illustrates a typical layout of beams for a sim ple saturated absorption apparatus Initially only a single beam passing through the cell is required which should be the full laser intensity for maximum sensitivity In this step one tunes the laser to an atomic transition and finds the optimum laser temperature current and mechanical ar rangement for stable operation When the cell is viewed through an IR viewer or a CCD television camera a strong track of fluorescence should become visible as the laser is tuned within the Doppler profile of an absorption line by mechanically rotating the grating It is helpful to ramp the PZT at a frequency of 20 Hz over a 15 V range during this search The diode current should be arbitrarily set between about 75 and 90 of I p If no fluorescence is apparent at any grating angle with the known tuning range the temptation to turn the grating farther or to ad just the vertical alignment of the grating should be resisted Most likely the laser has a tuning discontinuity that en compasses the desired wavelength The current sho
22. e use of optical feedback to control the laser frequency Reference 1 gives a lengthy technical review of the characteristics of laser diodes the use of optical feed back techniques to control them and various applications in atomic physics An earlier review by Camparo also gives much useful information primarily relating to free running diode lasers The use of a wavelength dispersive external cavity for diode laser tunning and mode selection was described by Ludeke and Harris and the spectral characteristics of external cavity stabilized diode lasers were investigated in detail by Fleming and Mooradian During the past several years our laboratory has carried out a large number of experiments in optical cooling and trapping and general laser spectroscopy of cesium and ru bidium using diode lasers In the course of this work we have developed a simple inexpensive design for a diode laser system that uses optical feedback from a diffraction grating This system produces over 10 mW of light with a bandwidth of well under 1 MHz and can be easily tuned over atomic resonance lines We now have over a dozen such laser systems operating including two in an under graduate teaching lab and the design has reached a rea sonable level of refinement There are many other designs for optical feedback systems and we make no claims for this one being superior However it is a reasonable com promise between several factors which are relevant
23. ent of wires outside the box will not transmit stress or vibration to the laser structure Finally one should make sure the enclosure is electrically grounded IV TEMPERATURE CONTROL Precise control of the temperature of both the baseplate and the diode laser itself is essential for the long term reliable operation of the laser at a particular wavelength We control these temperatures using identical independent servosystems The sensing element for the servo is a small thermistor which is part of a bridge circuit The amplified and filtered error signal drives a heater or thermoelectric cooler In this area of thermal control we have made the largest compromises of potential performance in order to simplify the mechanical and electrical designs Part of the reason we are willing to make this compromise is that we usually sense the output frequency of the laser and lock it directly to atomic transitions to insure long term stability at the sub MHz level This is discussed in Sec X The temperature of the diode laser mounting block is controlled only by heating which means that it must be kept 1 2 C hotter than the baseplate for proper tempera ture control The heating is done by a small 0 3 x 1 5 in adhesive film heater which is attached to the top or side of the laser mounting block The sensing thermistor rests in a small hole packed with heat sink compound in the mount K MacAdam A Steinbach and C Wieman 1102
24. g it on or off It is also wise to check that all the appropriate grounding connections have been made so that turning on and off nearby electrical equipment or static discharges do not cause current or voltage spikes that exceed the maximum allowed by the laser diodes When making these tests it is important to realize that lasers can be destroyed by spikes that last only a fraction of a micro second Only after the power supply has passed all these tests is it connected to the laser The cables from the power supply to the laser should be shielded and there should be no possibility of them being accidentally disconnected One has to be fairly careful in handling the lasers to avoid static discharges and it is a good idea to keep the leads shorted together as much as possible Normally such lasers come with handling instructions that should be fol lowed These instructions will usually also mention that a 1104 Am J Phys Vol 60 No 12 December 1992 fast reverse biased protection diode should be connected across the laser leads at the laser mounting block This diode protects against voltages spikes which may exceed the few volts of back bias a diode laser can tolerate We have found that the lifetime of diode lasers is substantially increased by also connecting several forward biased diodes across the leads at the same point as shown in Fig 9 These diodes have a large enough voltage drop that current does not flow through them under
25. ignment of the grating the output wavelength of the laser will be within about 2 nm of the wavelength specified by the manufacturer and near the center of the tuning range Small adjustments of the grat ing rotation screw vertical axis should smoothly shift the laser wavelength A region of the grating angle adjustment should be identified over which the laser can be tuned As one nears the end of the tuning range the laser output will be seen to hop back and forth or share power between two very different frequencies One is the fixed free running 1106 Am J Phys Vol 60 No 12 December 1992 frequency at which the laser will operate if there is too little or no feedback from the grating and the other is the angle dependent frequency set by the grating feedback At a given temperature tilting the grating should tune the out put wavelength over a range 10 to 30 nm depending on the particular laser and the amount of feedback Changing the diode temperature shifts the entire range by 0 25 nm C If the grating is misaligned the output wavelength will either be insensitive to small changes of the grating angle or will move only a small amount and then jump backwards Although this tuning may appear continuous when ob served on a low or medium resolution spectrometer there can actually be small gaps These occur because the wavelength dependent feedback of the grating dominates but does not always totally overwhelm feedback off
26. k box is connected to the other components Locking is not difficult after a little practice provided that the saturated absorption signals are not too noisy and the laser frequency jitter caused by environmental or electrical backgrounds is less than the saturated absorption line widths First the laser is tuned to the desired hyperfine multiplet of saturated absorption peaks and the ramp gain and ramp offset are adjusted both on the ramp generator and on the lock box so that one can zoom in to the desired side of a particular peak simply by turning down the ramp gain on the lock box to zero With feedback and output gain con trols set at minimum and the laser tuned to the side of a peak the error offset is adjusted to a value near O V as observed on an oscilloscope The feedback and output gains are then gradually increased until the circuit corrects for deviations from the desired lock point and thus holds the frequency on the side of the peak If the servolock seems to repel the saturated absorption peak the input invert switch is reversed to select the opposite slope When the laser is properly locked it should be possible to turn the feedback fully on and the output gain up to a point where the PZT begins to oscillate at about 1 kHz The best operating point is just below the onset of oscillation Lock ing is confirmed by noting that the setpoint indicated by the level of the now flat saturated absorption signal on the oscill
27. normal operation However if there is a large forward voltage these diodes turn on allowing the current to flow through them instead of the laser diode It may also be helpful to place a 10 0 current limiting resistor in series with the supply right at the laser diode at LD in Fig 9 and if modulation much above 1 MHz is not required ferrite beats on the supply lead at this point Switch S1 in Fig 8 should be used to short the D1 1N5711 D2 1N914 Fig 9 Protection diode wiring to laser K MacAdam A Steinbach and C Wieman 1104 supply to ground before connecting the laser and the cur rent control R3 should be fully off whenever S1 is tog gled VI ASSEMBLY AND TESTING A top view of the assembled laser is shown in Fig 1 A Diode mounting The laser diode with its protection diodes already wired on and its leads temporarily shorted together for safe han dling is mounted in its recess in the laser mounting block with the screws only gently tightened at first The desired orientation of the laser will produce a vertically polarized output beam and a widely diverging elliptical beam pattern whose major axis is horizontal This corresponds to the rectangular output facet of the diode chip having its longer dimension vertical Next the mounting block is attached to the baseplate by 4 40 screws extending from beneath The baseplate should then be secured to some temporary stand so that the uncol limate
28. oothly by use of the alignment jig later but vertical corrections require shimming first Note the vertical displacement of the spot from the aiming point A low spot will require raising the lens mount by about 0 0025 in per degree of misalignment and a high spot will require raising the diode block by the same amount Layers of aluminum foil avoiding crinkles or shim stock should be selected to shim the preliminary alignment beam height to within 1 of the aiming spot The alignment jig is installed next by screwing it to the laser baseplate using its oversize hole and a large washer or stack of washers so that it snugly touches the lens mount as shown in Fig 4 It should be positioned with its 2 64 screw and a spring or elastic cushion so that when the screws of the lens mount are released the mount can be pushed in both directions without losing contact With the lens mount screws now loosened the mount may be dis vlaced smoothly to bring the collimated spot horizontally to the aiming point A rubber band or finger pressure should be used to hold the loose lens mount against the jig A properly aligned laser will exhibit a symmetrical and elliptical beam spot The effects of aberration can be ob served by purposely misaligning the lens to one side or the other with the jig and a symmetrical behavior allows one to confirm that the designated aiming spot was initially correct After a satisfactory alignment and collimation has been o
29. oscope can be varied by the error offset control within a range from about 10 to 90 of the height of the selected peak without a noticeable change in the monitored error output Independently the error output can be varied over a wide range by the ramp offset control without af fecting the locked level of the saturated absorption signal When the laser is locked environmental noise appears on the error output and error signal monitor instead of on the saturated absorption signal because the error output compensates for laser frequency variations that would oth erwise occur The error signal monitor thus becomes an 1110 Am J Phys Vol 60 No 12 December 1992 excellent indicator of the magnitude and spectral charac teristics of the compensated noise out to the bandwidth of the servolock circuit The drift rate of the unlocked laser is normally under 5 MHz min when the system is properly stabilized and this slow drift is eliminated by locking The short term jitter amplitude of the unlocked laser frequency is typically 3 MHz on a l s time scale if the laser is on a reasonably stable lab table The short term intensity variations are much smaller than 1 When locked the laser frequency is stabilized to 1 MHz or better The locked diode laser described in this paper is well suited for studies of neutral atom cooling and trapping for which some elaborations of the servolock circuitry are de sirable A future paper will describe
30. ough not often necessary after stabilization require a delicate touch The laser structure takes one or more hours to fully stabilize inside its box with temperature control electronics active However prelimi nary output tests can proceed immediately if steady fre quency drift is not an obstacle Depending on the degree of stability required and the environment the laser may be operated on anything from an ordinary laboratory bench to a fully isolated optical table A room location near a load bearing wall or in a basement laboratory can often be worth the price of an expensive optical table Since the laser itself is one of the best vibration detectors obtainable experience will be the best guide IX SATURATED ABSORPTION SPECTROMETER The simplest spectroscopy one can perform with these lasers is to observe the absorption and Doppler free satu rated absorption spectra of rubidium or cesium This can easily be done in small glass vapor cells which are at room temperature Such spectroscopy experiments also provide the simplest way to determine the short and long term frequency stability and tuning behavior of the laser fre quency A Vapor cells Rubidium and cesium vapor cells can be obtained com mercially but they are usually rather expensive However they can be prepared quite easily if one has a vacuum pump and some basic glassblowing skills We use pyrex or quartz tubing typically 1 in in diameter and 2 to 4 in long
31. ower supplys are 7 C8 bypassed with 1 uF capacitors R32 o Q1 a D5 R35 2N2219 a 1N4001 gt 2420 MAX CURRENT TRIMMER 15U 12 R13 R14 180 199 MON OUT SpF be CE R30 t 33uF 8 06K R24 DPM gt OFFSET ADJ DIGITAL PANEL METER l Cle t C11 Re eee F T 75 u 4 uF Fig 8 Laser current control circuit V LASER DRIVE ELECTRONICS The circuit diagram for the laser current controller is shown in Fig 8 This is a stable low noise current source The output current can be modulated rapidly by sending a voltage into the RF MOD IN input If such modulation is not needed and novices may be operating the laser it is wise to disconnect or cover this input to minimize the possibility of accidentally damaging the laser The output current of the supply is limited by potentiometer R36 to a value that cannot exceed the maximum allowed for the diode laser The primary concern when working with the current source is to avoid damaging the laser with an unwanted current or voltage spike In Ref 1 we discuss this danger at some length so here we will just provide a few helpful techniques To avoid accidents we always carefully test a new power supply with resistors and light emitting diodes in place of the laser We check that it produces the voltage and current desired and that there are no significant tran sients when turnin
32. plastic or glass When the two photodiodes are properly positioned the differential output signal cancels the large and feature less Doppler profile of the absorption line and allows sat urated absorption features from the first probe beam to appear on a nearly flat background If the Doppler broad ened absorption is observed but the saturated absorption 1109 Am J Phys Vol 60 No 12 December 1992 peaks cannot be seen it often means that there is too much background gas in the vapor cell E Saturated absorption patterns in Rb and Cs After saturated absorption peaks have been observed one can compare the patterns to known hyperfine struc tures of the ground and excited states to assess the electri cal tuning range possible without hopping external cavity modes and to establish the tuning direction The widths and resolution of the saturated absorption peaks for a given resonance line will depend on electronic time constants the triangle wave frequency and possibly on diode current in addition to alignment and intensity factors noted above Figure 13 shows several saturated absorption patterns in Rb 780 nm and Cs 852 nm vapors photographed from an oscilloscope These may aid new users in finding their way Note that the patterns contain both true Doppler free peaks and crossover peaks which occur at frequencies v 2 for each pair of true peaks at frequency v and v The crossovers are often more intense than the
33. rief summary of what must be specified and our recommendations for suppliers The basic requirement for a diode laser which is to be used in this system is that it have a high reflectivity coating on the back facet and a reduced reflectivity on the front or output facet Very inexpensive diodes which produce a few milliwatts of power have two uncoated facets and will not work very well We have used 20 mW lasers but their performance is marginal However we have found that any laser we have tried that is specified to provide 30 mW or more single mode will have the necessary coatings and will work well It will provide narrowband laser light that is tuneable over 20 30 nm If one wants this range to cover the 852 nm cesium or 780 nm rubidium resonance lines the diode laser wavelength must be specified when pur chasing This greatly complicates the purchasing We have tried numerous suppliers but have now settled on STC as our supplier of lasers for 852 nm and Sharp as the supplier for 780 nm The Sharp lasers are far less expensive and can usually be obtained rather quickly since 780 nm is near the center of the distribution of their normal mass produced product This is not the case for 852 nm and thus the lasers must be produced as a custom run STC has made several such custom runs and hence usually has 852 nm lasers available although they cost 3 to 4 times more than the Sharp lasers The long wavelength edge of the distri bution of Sh
34. rlapped probe beam The counter propagating saturating beam can easily be aligned to overlap the probe beam at a 1 intersection angle or smaller The intensities of the beams are not important for initial adjustments but typically only a small fraction of the laser output less than a few percent should be used for the saturated absorption Reflection from a microscope slide provides an ample intensity that will allow further attenuation by neutral density filters or exposed photo graphic film When adequate pump and probe beam over lap has been obtained small saturated absorption dips should become evident near the center of the absorption line Fig 12 They may be recognized unambiguously by their disappearance from the Doppler profile if the satu rating beam is blocked The height of the narrow dips may be maximized by adjusting the alignment The width can be reduced by reducing the angle of intersection of the overlapped beams and by attenuating either or both beams to avoid power broadening The triangle wave amplitude and dc offset can be adjusted to zoom in on a particular region of the scan For more detailed observations it is helpful to unblock the second probe beam This second probe beam is directed into a photodiode identical to the first and wired in parallel with reversed polarity The two probe beams can easily be obtained by utilizing the reflections off both front and rear surfaces of a piece of 3 8 in thick transparent
35. scount department stores For an adequate model prices for a camera lens and monitor will range from 500 to over 1000 6 Sharp Diode Laser LTO25MDO 170 85 wavelength 780 nm Added Value Electronic Distributors Inc local Sharp distributor 4090 Youngfield Street Wheatridge CO 80033 303 422 1701 STC LTS0A 034 laser diodes STC was recently pur chased by Northern Telecom wavelength 852 nm We have purchased these lasers for 650 from a German distributor Laser 200 GMBH Argelsrieder Feld 14 D 8031 Werling Germany 7 Minco Thermofoil Kapton Heater Minco 8941 P N HK5207R12 5L12A 23 50 10 PSA Pressure sensitive K MacAdam A Steinbach and C Wieman 1110 adhesive sheet 4 00 Minco Products Inc 7300 Com merce Lane Minneapolis MN 55432 612 571 3121 x3177 8 Diffraction grating 1200 1 mm 500 nm blaze P N C43 005 72 85 750 nm blaze P N C43 210 72 85 Ed mund Scientific address as above 9 Thermistor P N 121 503JAJ QO1 8 25 Fenwall Electronics 450 Fortune Blvd Milford MA 01757 also available from electronics distributors 10 PZT disk P N PE 8 0 75 Murata Erie 7BB 27 4 All Electronics Corp P O Box 567 Van Nuys CA 91408 818 904 0524 11 Kinematic Mirror Mount Mod MML 52 00 Thorlabs Inc P O Box 366 Newton NJ 07860 201 579 7227 12 Fine Adjustment Screw Mod AJS 0 5 30 00 Newport Corp P O Box 8020 18235 Mt Baldy Circle Fountain Valley CA
36. t from the accumulated practical experience in another laboratory We will first discuss the construction or purchase of the basic components and then explain how to put them to gether align the laser system and tune the frequency Fi nally we discuss how to observe saturated absorption spec tra and how to use these spectra to evaluate the laser performance or to actively stabilize the laser frequency by locking it to narrow saturated absorption features 1992 American Association of Physics Teachers 1098 Fig 1 Assembly top view of laser The arrow showing the blaze direction on the grating is for the low feedback large output case II SYNOPSIS OF COMPONENTS As shown in Fig 1 the laser system has three basic components a commercial diode laser a collimating lens and a diffraction grating These components are mounted on a baseplate The laser and lens are mounted so that the lens can be carefully positioned relative to the laser to insure proper collimation The diffraction grating is mounted in a Littrow configuration so that the light dif fracted into the first order returns to the laser As such the grating serves as one end mirror of a laser cavity with the back facet of the diode providing the second mirror This means the grating must be carefully aligned and very stable To achieve this we mount the grating on a standard commercial mirror mount which is attached to the base
37. t heat up enough to cause thermal runaway of the TEC The base plate temperature is monitored using a thermistor glued onto the middle of the baseplate Since the thermal time constant for the baseplate is much longer some adjustment or removal of capacitors C1 C2 and C3 from the tem perature control circuit may be desirable to improve sta bility If the laser is cooled below the dew point condensa tion may form This may be avoided by flushing gently with dry N An alternative to controlling the baseplate temperature is to control the temperature of the entire enclosure This is more effort because much more heating or cooling power is needed and the thermal time constant is very long We find however that this technique gives better ultimate sta bility of the laser alignment For most purposes we have found that this is not worth the effort However the small additional effort required in putting insulation on the out side of the aluminum enclosure to attenuate room temper ature fluctuations is worthwhile K MacAdam A Steinbach and C Wieman 1103 R9 CURRENT CONTROL MON 150K Ri R2 CURRENT 1 SK 2K CONTROL T154 o nN AW 1 Ci R3 LM399 luF D1 i i NOTES 1 The digital panel meter s an DG 1N4001 Acculex DP 650 meter 2 R The blocks on the bases of 2 15U Ql and Q2 are ferrite beads WN 3 Resistors are 1 metal film 4 All op amps p
38. tage is applied the piezo electric stress causes the backing to dish on the opposite side Two such elements can be attached back to back doubling the displacement of a single one by lightly sol dering the adjacent brass backings at four places around their circumference If necessary for clearance in the grat ing mount some of the excess brass can be clipped away without damaging the piezoelectric center The double PZT is wired by lightly soldering one connection to the brass and the other to the two silver plated piezo elements in parallel For this and all other wiring of the laser it is best to select a limp insulated wire that will not transmit vibration to the laser structure Rubber covered No 24 test prod wire has been found suitable After the PZT is assem bled and wired and the grating is glued to the mount the 1102 Am J Phys Vol 60 No 12 December 1992 PZT should be inserted between the mounting plate and the ball end of the adjusting screw as shown in Fig 6 Small pieces of mylar should be inserted to electrically isolate the PZT from the mount The PZT will provide about 1 um of displacement when 15 V are applied F Enclosure for the laser An aluminum enclosure should be fabricated to hold the laser It should have a sufficient thermal mass and conduc tivity to aid in temperature stabilization Such a box can be made out of rectangular side plates screwed together placed on a rectangular baseplate
39. the AR coated output facet of the chip If it proves impossible to excite some desired atomic absorption line by tilting the grating it is necessary to operate at a different temperature and or current This is best assessed by means of an atomic absorption cell discussed below since the gaps in tuning can be narrow and vary randomly from one laser to an other It is helpful to record the tuning rate vs grating rotation about 14 nm turn with an 80 thread per in screw pitch because one can easily mistune the grating grossly requiring a retreat to earlier steps in the alignment process After the grating rotation has been set to produce approximately the correct wavelength the vertical align ment should be rechecked using the threshold current tech nique The simple laser design described here suffers from a defect that may be annoying in wideband usage Its output beam is deflected horizontally as the wavelength is scanned approximately at the angular rate dOBEAM d for grating constant d This is normally of no consequence however for saturated absorption or neutral atom trap ping e g in Rb where the 5s 5p3 hyperfine multiplets of the two naturally occurring isotopes span a total of less than 0 014 nm If it is necessary to avoid beam deflection the simplest technique is to take the output beam off a beam splitter inserted between the collimating lens and the grating d A 2 0 08 deg nm VI
40. the coated surface of the grating into pieces of the desired size A 1X1 in grating will yield six suitable pieces Saw the grating in an abrasive wheel glass saw by holding the support block on its edge as the saw cuts di rectly into the face of the grating Make sure the saw cuts penetrate completely through the grating into the support block without severing the block Then melt off the cut segments The nail polish can then be removed by sub merging the grating segments in a small beaker of metha nol and placing the beaker in an ultrasonic cieaner Re move the gratings with tweezers being very careful to avoid any contact with the now exposed ruling surface refill the beaker with fresh methanol and repeat once or twice until the gratings when drained and dried appear completely clean Harsher solvents may attack the plastic substrate of replica gratings but methanol has been found to be safe and effective After cleaning the grating is attached to the movable face of the grating mount in a location where the colli mated laser beam will strike near the middle of the grating Care should be taken to make the rulings vertical A stiff but readily removable adhesive such as Duco cement is recommended for attaching the grating to the mount The grating segment can be easily damaged when it is necessary to remove it or shift its position unless it can be detached with little physical force If necessary the efficiency of most inexpensive
41. the correct height The collimating lens should be firmly fastened into its mounting block with a set screw and the mounting block should be loosely screwed to the laser baseplate with the flat side of the lens toward the diode The precision adjusting screw that pushes against the baseplate hinge should be advanced so that the hinge is opened enough to allow plus and minus 0 020 in of motion without losing contact with the ball end of the screw Next insert a clean microscope slide about 1 mm thickness between the lens flange and the front face of the diode laser mounting block While holding the lens block the slide and the diode block together with finger pressure observe the beam spot again with the IR viewer It should be pos sible to slide the lens block back and forth to bring a more concentrated intensity maximum near to the original aim 1105 Am J Phys Vol 60 No 12 December 1992 ing point of the laser Temporarily tighten the screws that hold the lens block in that position and remove the slide Next adjust the precision screw to bring the laser beam to a sharp focus on the wall By a very slight adjustment of the screw the beam should then be brought to collimation in an oblong spot about 5 mm wide It should be confirmed that no focus occurs between the laser and the wall This constitutes a preliminary alignment The beam spot will very likely fall 2 or more from the aiming spot Horizontal corrections can be made sm
42. the laser frequency and changes in the cavity alignment The baseplate is either heated or cooled depending on the requirement Heating is much simpler since it only requires the attaching of a film heater to the baseplate The film heater is similar to that used on the laser except it is larger in area and power output To keep the baseplate controlled it is necessary that it be at least 1 2 C above the room temperature and the laser must be an equal amount hotter than the baseplate This is not difficult if the laser s free running wavelength at room temperature is shorter than wavelength desired In that case it is advantageous to heat the laser If however the laser s free running wavelength is significantly to the red 1103 Am J Phys Vol 60 No 12 December 1992 the laser should be run near or below room temperature In this case the baseplate must be cooled below room tem perature using a thermoelectric cooler TEC This is somewhat more trouble and the vibration isolation pads between the baseplate and the bottom of the enclosure are now replaced by a rigid TEC The TEC is a square 1 5 in on a side and fits between the baseplate and the aluminum plate which is the bottom of the enclosure A thin layer of heat sink compound is applied on both sides of the TEC to insure good thermal contact The bottom plate of the en closure must have a large enough surface area or be in contact with a thermal reservoir so that it does no
43. to 1098 Am J Phys 60 12 December 1992 many laboratories 1 low cost about 400 not including labor 2 ease of construction several of these systems have been built by novice undergraduates and 3 reli ability These lasers have achieved several notable suc cesses in experiments on cooling and trapping cesium at oms and the design has been successfully duplicated in a number of other laboratories We prepared this article in response to a large number of requests for detailed instruc tions on how to build and operate such a system This article provides a detailed and fully comprehensive recipe for construction of the system and its use to observe satu rated absorption spectra in a rubidium or cesium vapor cell We refer the reader to Ref 1 and the references therein for information about the physics of laser diodes and the factors that motivated this design as well as design alternatives In this paper we have attempted to respond to three frequent requests for information we receive The first is from the undergraduate wanting to do high resolution la ser spectroscopy for a project without expert local super vision The second is from the faculty member who wants to construct a teaching laboratory experiment and wants instructions that can be given to a technician or undergrad uate with favorable results The third is from the research scientist who wants to use diode lasers in an experiment and would like to benefi
44. trapping of Rb and Cs atoms from a vapor cell in a user manual style similar to that used here ACKNOWLEDGMENTS This work was supported by the NSF and ONR We are indebted to many people who contributed ideas which have been incorporated into the present design Much of the basic design work was carried out by Bill Swann Kurt Gibble and Pat Masterson Steve Swartz Jan Hall and nearly every member of the Wieman group during the past several years have also provided valuable contributions Melles Griot Inc loaned us an excellent diode laser current supply which was used for part of this work APPENDIX PARTS AND SUPPLIERS 1 Collimating lens 1403 108 75 00 f 5 mm nu merical aperture 0 5 Rodenstock Precision Optics Inc 4845 Colt Road Rockford IL 61109 815 874 8300 2 Sorbothane Pad P N C37 000 49 95 Edmund Sci entific 101 E Gloucester Pike Barrington NJ 08007 1380 609 573 6250 3 Photodiode Pin 10D 1 cm active area 55 25 United Detector Technology Sensors 12525 Chadron Avenue Hawthorne CA 90250 213 978 1150 x360 4 Kodak IR detection card R11 236 49 50 Edmund Scientific address as above 5 Hand held infrared viewer P N 84499 1195 00 FJW Optical Systems Inc 629 S Vermont Street Pala tine IL 60067 6949 708 358 2500 A less expensive alternative is to use a CCD surveillance camera These can be purchased from many sources in cluding home and office security companies and di
45. true peaks F Typical tuning rates observed by saturated absorption Tuning rates for the grating feedback laser operated with a single longitudinal mode of the external cavity de pend on geometrical thermal and electrical properties of the laser components In particular tilting the grating changes both the wavelength of light diffracted back to the diode and the length of the cavity These two effects inter act in determining the change of output frequency Typical tuning rates for a 780 nm laser having a 20 mm cavity on an aluminum baseplate are 1 diode drive current 200 MHz mA 2 temperature change of diode 4 GHz C 3 temperature change of baseplate cavity length 7 GHz C 4 grating angle change 80 pitch screw 5 x 10 MHz turn and 5 piezoelectric tuning 1 GHz V K MacAdam A Steinbach and C Wieman 1109 PHOTODIODE ABSORPTION CELL CURRENT SUPPLY TEMPERATURE eet yee IOSCILLOSCOPE CONTROL A BOX RAMP Fig 15 Electronic layout schematic For operation without the servolock box the ramp is connected directly to the PZT as shown by the dashed line X SERVOLOCKED OPERATION OF THE DIODE LASER For stabilized operation of the laser it may be locked to either side of any of the sufficiently well resolved saturated absorption peaks such as those shown in Fig 12 A simple servolock circuit is given in Fig 14 Figure 15 indicates how the loc
46. uld be changed several mA and the procedure repeated If this still fails the temperature should be changed up or down 0 5 C to 1 C and the search for the absorption line should be repeated If this process is iterated several times without success it may be desirable to look once again with the grating spectrometer to confirm that the laser is still tuning in the desired range and that the grating has not been grossly misaligned by a random walk When one finds a grating position which produces fluorescence the current can be adjusted to maximize the fluorescence K MacAdam A Steinbach and C Wieman 1107 PHOTODIODE CURRENT 85 Rb F 3 gt F PZT VOLTAGE Fig 12 Single beam saturated absorption in Rb F 3 F Once a proper temperature has been set it should not be necessary to change it However when the laser is turned on in the morning it is not uncommon to find that the proper drive current has changed by up to 1 mA or at a fixed current minute readjustment of the grating angle is needed to hit the absorption line again This drift may be caused by environmental changes hysteresis in the electri cal tuning characteristics aging of the diode or mechani cal creep of laser cavity components C Piezoelectric scanning After the laser is mechanically tuned onto an absorption line as observed in the IR viewer the transmitted probe beam should be attenuated so that the intensity is less than 3 mW cm
47. ximum back into the laser The output beam is the zero order beam or specular reflection maximum which passes horizontally beside the collimator block and out of the enclosure The direction of the blaze is toward the output beam A laser diode whose free running wavelength is within about 3 nm of the desired wavelength requires less feedback for stabi lized operation than a laser that must be pulled more se verely For this case lower diffraction efficiency and thus a shorter blaze wavelength 500 nm is suitable and this allows more power to be brought out in the zero order beam If a laser must be pulled more severely a longer blaze wavelength 750 nm is used to provide stronger feedback at the price of lower output power When a grating of suitable blaze has been selected it may be cut down to a small size since only about 0 3 in parallel to the rulings and 0 5 in perpendicular is required Thus several gratings can be had for the price of one and the others may be used to duplicate the diode laser system or for testing grating properties outside the laser The cutting may be safely done as follows Apply a generous coating of clear acetate fingernail polish to the ruled face of the grat ing Spread the fluid using a soft camel s hair brush and avoid physical contact with the grating After the coating is thoroughly dry wax the back of the grating to a block of bakelite or phenolic to support the grating while it is sawed Mark
48. xternal cavity mode to another These mode hops may be as far as 8 GHz apart but will exhibit somewhat random spacings as well as hysteresis 1108 Am J Phys Vol 60 No 12 December 1992 87Rb F 1 gt F SATURATED ABSORPTION SIGNAL RELATIVE 87Rb F 2 gt F 0 100 200 300 400 500 a FREQUENCY MHz m m 133C5 F 3 gt F 133Cs F 4 gt F SATURATED ABSORPTION SIGNAL RELATIVE 0 200 400 600 800 b FREQUENCY MHz Fig 13 Saturated absorption curves for a Rb and b Cs The Rb F 2 F peaks are broader than the others because they were made with a different setup The widths and relative heights are affected by beam alignment beam intensities electronic damping constants and absorption cell pressure These are only representative results K MacAdam A Steinbach and C Wieman 1108 SET RS POINT 29K 10T 6 95V All capacitor values are in uF FEEDBACK ERROR SIG MONITOR R6 3M PZT OUT OUTPUT 15u RAMP IN R10 R11 pean ET Rig OUTPUT SOK 20K OFFSET RAMP GAIN 15U Fig 14 Servolock circuit D Observing the saturated absorption The full saturated absorption setup of Fig 10 is required for a more detailed test of stability and tuning rates and for locking of the laser output frequency at the level of 1 MHz or better When first observing a saturated absorption sig nal it is useful to block the nonove
Download Pdf Manuals
Related Search