U.S. patent application number 11/612465 was filed with the patent office on 2007-07-05 for manufacturable vertical extended cavity surface emitting laser arrays.
Invention is credited to Rene Dato, John Green, Michael Jansen, Dicky Lee, Aram Mooradian, Andrei V. Shchegrov, Arvydas Umbrasas, Jason P. Watson.
Application Number | 20070153866 11/612465 |
Document ID | / |
Family ID | 39230898 |
Filed Date | 2007-07-05 |
United States Patent
Application |
20070153866 |
Kind Code |
A1 |
Shchegrov; Andrei V. ; et
al. |
July 5, 2007 |
MANUFACTURABLE VERTICAL EXTENDED CAVITY SURFACE EMITTING LASER
ARRAYS
Abstract
Arrays of vertical extended cavity surface emitting lasers
(VECSELs) are disclosed. The functionality of two or more
conventional optical components are combined into an optical unit
to reduce the number of components that must be aligned during
packaging.
Inventors: |
Shchegrov; Andrei V.;
(Campbell, CA) ; Watson; Jason P.; (San Jose,
CA) ; Lee; Dicky; (Santa Clara, CA) ;
Umbrasas; Arvydas; (Cupertino, CA) ; Dato; Rene;
(Pleasanton, CA) ; Green; John; (Scotts Valley,
CA) ; Jansen; Michael; (Palo Alto, CA) ;
Mooradian; Aram; (Kentfield, CA) |
Correspondence
Address: |
COOLEY GODWARD KRONISH LLP
3000 EL CAMINO REAL
5 PALO ALTO SQUARE
PALO ALTO
CA
94306
US
|
Family ID: |
39230898 |
Appl. No.: |
11/612465 |
Filed: |
December 18, 2006 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
11396341 |
Mar 30, 2006 |
|
|
|
11612465 |
Dec 18, 2006 |
|
|
|
11193317 |
Jul 29, 2005 |
|
|
|
11396341 |
Mar 30, 2006 |
|
|
|
60592890 |
Jul 30, 2004 |
|
|
|
60667201 |
Mar 30, 2005 |
|
|
|
60667202 |
Mar 30, 2005 |
|
|
|
60666826 |
Mar 30, 2005 |
|
|
|
60646072 |
Jan 21, 2005 |
|
|
|
60689582 |
Jun 10, 2005 |
|
|
|
Current U.S.
Class: |
372/50.124 ;
348/E9.027; 372/22; 372/50.11; 372/99 |
Current CPC
Class: |
H01S 5/18308 20130101;
H04N 9/3129 20130101; H01S 5/0234 20210101; H01S 3/109 20130101;
H01S 3/1062 20130101; H01S 5/18305 20130101; H01S 5/18358 20130101;
G02B 27/286 20130101; H04N 9/3111 20130101; H01S 5/02326 20210101;
G03B 21/2033 20130101; H01S 5/4093 20130101; G02B 27/141 20130101;
H01S 5/423 20130101; H01S 5/0071 20130101; H01S 5/0217 20130101;
H01S 5/141 20130101; H01S 5/18388 20130101; H04N 9/315 20130101;
G02B 27/48 20130101; G02B 27/102 20130101; H01S 3/08072 20130101;
H04N 9/3114 20130101 |
Class at
Publication: |
372/050.124 ;
372/050.11; 372/099; 372/022 |
International
Class: |
H01S 3/10 20060101
H01S003/10; H01S 5/00 20060101 H01S005/00; H01S 3/08 20060101
H01S003/08 |
Claims
1. A semiconductor, vertical, extended-cavity, surface-emitting
laser array with intra-cavity nonlinear frequency doubling,
comprising: an electrically pumped, surface-emitting semiconductor
gain chip having an array of emitters formed in a common
semiconductor substrate, each emitter generating light about a
fundamental wavelength; an end reflector spaced apart from said
surface-emitting semiconductor gain chip to define an extended
cavity for each of said array of emitters, said end reflector
having a high reflectivity at the fundamental laser wavelength; a
nonlinear crystal disposed within the extended cavity to provide
frequency doubling of light for the entire array of emitters; a
wavelength control element disposed within the extended cavity to
provide wavelength control for the entire array of emitters; and a
polarization control element disposed within the extended cavity to
provide polarization control for the entire array of emitters; said
surface-emitting laser array having an optical unit having the
functionality of at least two of said gain chip, said end
reflector, said nonlinear crystal, said wavelength control element,
and said polarization control element to reduce the number of
components that must be aligned during packaging.
2. The laser array of claim 1 wherein said end reflector is a
volume Bragg grating, said volume Bragg grating serving said end
reflector and said wavelength control element.
3. The laser array of claim 2 wherein the volume Bragg grating
includes a high reflectivity optical coating deposited at an outer
face of said volume Bragg grating to increase the circulating power
intensity in the laser cavity at the fundamental wavelength and
increase nonlinear conversion efficiency.
4. The laser array of claim 2, wherein said nonlinear crystal is
attached to said volume Bragg grating as an unit serving as said
end reflector, wavelength control element, and nonlinear
crystal.
5. The laser array of claim 2, wherein said unit further includes a
polarization control element attached to said nonlinear crystal
such that said unit serves as said end reflector, wavelength
control element, nonlinear crystal, and polarization control
element.
6. The laser array of claim 5, wherein said polarization control
element is a waveplate.
7. The laser array of claim 6, wherein said polarization control
element is a wire grid polarizer.
8. The laser array of claim 1 wherein the wavelength control
element is a thin-film coating interference filter
9. The laser array of claim 1 wherein the polarization control
element is a coated beamsplitter providing loss for an undesirable
polarization and substantially no loss for a desirable polarization
of light at said fundamental frequency.
10. The laser array of claim 1 wherein the polarization control
element is a planar element attached to another planar component to
form a planar sub-unit.
11. The laser array of claim 10, wherein said polarization control
element is a wire-grid polarizer.
12. The laser array of claim 1 wherein the polarization control
element is a birefringent crystal, which may be separate or the
same as the nonlinear crystal, providing walk-off and misalignment
for the undesirable polarization
13. The laser array of claim 1, wherein a planar array of lenses is
disposed in said extended cavity to provide transverse mode control
for said array of lasers.
14. The laser array of claim 13 wherein the lenses are thermal
lenses associated with each emitter of the surface-emitting
array
15. The laser array of claim 1 wherein a planar array of apertures
is disposed in the extended cavity.
16. The laser array of claim 15 wherein the planar array of
apertures are lithographically integrated with the gain chip.
17. The laser array of claim 1, wherein said polarization control
element comprises a dichroically coated beamsplitter, said
beamsplitter coupling light at a second harmonic frequency out of
said extended cavity.
18. The laser array of claim 1, wherein said cavity has a forward
direction for beams traversing towards said end reflector and a
backwards direction for beams traversing towards said gain chip,
said beamsplitter receiving and re-directing second-harmonic beams
traversing in the backward direction outside of the laser
cavity.
19. The laser array of claim 19 further comprising a corner-turning
mirror disposed outside of said extended cavity to direct an array
of backward-propagating second-harmonic beams into parallel
paths.
20. The laser array of claim 19 wherein said backward propagating
second harmonic are reflected and in the same direction as second
harmonic beams traveling in a forward direction coupled out of said
extended cavity via said end reflector.
21. The laser array of claim 20 wherein the forward-propagating
generated second-harmonic beams are reflected back into the cavity
by a coating on the nonlinear crystal of the volume Bragg grating
either along the original path or at an angle using a tilt or wedge
and then re-collected by the dichroic beamsplitter together with
the backward-propagating generated second-harmonic beams.
22. The laser array of claim 1 wherein a waveplate is disposed
within the extended cavity to rotate the polarization of the
backward-propagating generated second-harmonic beams which are then
back-reflected onto their original path by a coating on the
waveplate or the surface-emitting laser chip and then re-combined
with the forward-propagating generated second-harmonic beams of the
orthogonal polarization.
23. The laser array of claim 23 wherein the waveplate, the
nonlinear crystal, and the volume Bragg grating are monolithically
bonded to form a low-cost subunit defining the extended cavity.
24. The laser array of claim 1 wherein a waveplate is disposed
within the extended cavity after the nonlinear crystal to rotate
the polarization of the forward-propagating generated
second-harmonic beams which are then back-reflected onto their
original path by a coating on the waveplate or the end reflector
such as a volume Bragg grating and then re-combined with the
backward-propagating generated second-harmonic beams of the
orthogonal polarization.
25. The laser array of claim 1 wherein all the optical elements in
the extended cavity are mechanically aligned using passive
alignment to a high-precision fiducial marks of the laser
package.
26. The laser array of claim 1 wherein one of the optical elements
is wedged or tilted and coated to provide a reflection and angular
separation for the forward- and backward-propagating
second-harmonic beams.
27. The laser array of claim 1 wherein an array of dome lenses is
formed on the semiconductor surface-emitting chip to help optimize
the spatial mode of the laser cavity and to focus the fundamental
beams into the nonlinear crystal.
28. The laser array of claim 27 wherein the array of domes is
coated with an anti-reflective coating at the fundamental
wavelength and with a high-reflective coating at the
second-harmonic wavelength to provide reflection of the
backward-propagating second-harmonic beam which is also spatially
expanded and thereby has a reduced spatial overlap with the
forward-propagating second-harmonic beam.
29. The laser array of claim 1 in which the output array of
second-harmonic beams is re-shaped by a diffractive optical element
to achieve the desired illumination distribution.
30. The laser array of claim 1 in which speckle is reduced by the
selection of the number of emitters and by promoting
multi-longitudinal mode operation to decrease the overall spatial
and spectral coherence of the laser light source.
31. The laser array of claim 1 in which the forward- and
backward-propagating second-harmonic beams are recombined through
an optical paths separating them by more than the coherence length
of the laser to reduce interference between these beams.
32. A laser array as in claim 1 in which one or more of the
elements are combined, and the remaining elements are packaged in
such a fashion that the total volume of the package is less than
one cubic inch.
33. The laser array of claim 32 in which the elements are combined
and packaged in such a way that the total volume of the package is
less than two cubic inches.
34. A semiconductor, vertical, extended-cavity, surface-emitting
laser array with intra-cavity nonlinear frequency doubling
comprising an electrically pumped, surface-emitting semiconductor
gain chip having an array of emitters formed in a single
semiconductor substrate, each emitter generating light about a
fundamental wavelength; a volume Bragg grating spaced apart from
said surface-emitting semiconductor gain chip to form an end
reflector defining an extended cavity for each of said array of
emitters, said volume Bragg grating having a high reflectivity at
the fundamental laser wavelength and providing spectral filtering
to control the frequency of each emitter of said laser array; and a
nonlinear crystal to provide intra-cavity frequency conversion of
light for the entire array of emitters.
35. The laser array of claim 34, wherein said laser array is
configured to require only one critical optical alignment in which
said volume Bragg grating is packaged in alignment with said gain
chip and other components do not require a critical optical
alignment.
36. The surface-emitting laser array of claim 34, further
comprising a waveplate attached to said nonlinear crystal.
37. The surface emitting laser of claim 34, wherein said nonlinear
crystal is attached to said volume Bragg grating.
38. The surface emitting laser of claim 34, wherein lenses for
controlling the spatial mode of each emitter are formed in said
gain chip.
39. The surface emitting laser of claim 38, wherein said lenses
comprise thermal lenses.
40. The surface emitting laser of claim 38, wherein said lenses
comprise an array of lenses bonded to said gain chip.
41. The surface emitting laser of claim 34, further comprising an
array of apertures formed on said gain chip for controlling a
spatial mode.
42. The surface emitting laser of claim 34, further comprising a
polarizing element formed on another wavelength control element
disposed within the extended cavity.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation-in-part of U.S. Pat. Ser.
No. 11/193,317, "Projection Display Apparatus, System, and Method,"
which claims the benefit and priority to provisional application
Nos.: 60/592,890, filed on Jul. 30, 2004; 60/667,201 filed on Mar.
30, 2005; 60/667,202 filed on Mar. 30, 2005; 60/666,826 filed on
Mar. 30, 2005; 60/646,072 filed on Jan. 21, 2005; and 60/689,582
filed on Jun. 10, 2005, the contents of each of which are also
hereby incorporated by reference.
FIELD OF THE INVENTION
[0002] The present invention is generally related to light sources
for use in display systems. More particularly, the present
invention is directed to utilizing semiconductor lasers to replace
conventional white light discharge lamps in projection display
systems.
BACKGROUND OF THE INVENTION
[0003] Digital light processing (DLP) is of interest for projection
display systems, such as projecting images in conference rooms,
home television systems, advertising displays, automobile dashboard
and heads-up displays and other applications. The light engine of a
DLP system typically includes the light source and other components
required to generate light at several different colors. Individual
color components are spatially modulated to generate individual
pixels having selected color intensities. Additional optical
elements focus the light onto a display.
[0004] A DLP system typically includes a spatial light modulator
that modulates a light source in order to generate pixels at a
projection surface with controlled intensity. A light valve is a
type of spatial light modulator that modulates light across an
array of elements. A light valve typically modulates optical
transmission or reflection properties across an array. For example,
a reflective liquid crystal light valve utilizes an array of liquid
crystals elements to modulate the intensity of reflected light
across the array. Another common type of light valve is a digital
micro-mirror device chip (often known as a "digital light valve")
that has an array of movable micro-mirrors that can be individually
tilted between two positions to vary the amount of light per-pixel
that is reflected onto a display surface. Digital mirror devices
can switch fast enough to allow a single spatial modulator to be
used in a projection system operating in a color sequential mode.
This can provide cost savings over per-color modulator designs with
slow spatial modulators.
[0005] Conventionally a bright white light is used as the light
source for digital projection systems. For example, the bright
white light source is often implemented using an Ultra High
Pressure (UHP) arc discharge lamp, which is a compact white light
source with a very high luminance that was developed by Philips
Electronics. A rotating color wheel is used to separate out red,
green, and blue light from the white light source. Thus, when the
red filter of the color wheel is aligned to the white light source,
red light is focused onto the spatial modulator for the red color
of pixels, and so one for the green and blue filters of the color
wheel.
[0006] There are several drawbacks to conventional DLP systems.
First, the image is sometimes not as bright as desired.
Conventional white light sources produce a limited number of lumens
of light. Additionally, conventional DLP systems waste a
considerable amount of the light energy. Second, some attributes of
the displayed image, such as color saturation, are deleteriously
affected by the color wheel, which can introduce artifacts into the
displayed image. Third, DLP systems include expensive optical
elements.
[0007] Light emitting diodes (LEDs) are one alternative to white
light sources. However, conventional LED light sources tend to be
more expensive than UHP lamps. Additionally the brightness and the
number of lumens that can be coupled to a display screen is
typically about a factor of two lower for LEDs compared with UHP
lamps. As a result, LEDs have many limitations as light sources in
projection display systems
[0008] Semiconductor lasers have a number of potential advantages
as light sources in display systems. Semiconductor lasers have
high-brightness, low etendue, extended color gamut, and the
capability for modulation. For example, several discrete lasers of
different colors can be packaged to generate light at different
colors.
[0009] However, prior art semiconductor lasers have several
drawbacks as light sources for display systems. Compared with UHP
white light sources, conventional semiconductor lasers are not
cost-competitive and have a lower power (i.e., smaller total number
of lumens of light). Additionally, semiconductor lasers typically
have unacceptable speckle characteristics due to the high coherence
of semiconductor lasers. In the context of a display system, a high
degree of speckle results in light and dark patches across an image
due to constructive and destructive interference from scatter
centers.
[0010] In the prior art it was known that semiconductor lasers were
not cost competitive with UHP lamps in many projection display
applications. For example, for rear-projection televisions (RPTV)
it was known that the light source must be able to provide 300 to
600 lumens of light for each color at a total cost of no more than
about $100. See, e.g., K. Kincade, "Optoelectronics Applications:
Projection Displays: Laser based projector target consumer market,"
Laser Focus World, December 2005, the contents of which are hereby
incorporated by reference. For a laser based system, 300 to 600
lumens corresponds to about 3 to 5 Watts for each color. However,
in the prior art commercially available semiconductor lasers having
the requisite luminance and satisfactory beam properties could not
meet the total price point of $100 required for a RPTV system.
[0011] The cost of visible semiconductor lasers depends upon many
factors. Nonlinear frequency conversion process may be used to
generate red, green, and blue (RGB) colors. However, traditional
approaches result in a complex system that is difficult to
manufacture. Conventional visible high power semiconductor lasers
require a variety of optical elements to maintain wavelength
control, polarization control, and provide frequency conversion of
a pump light source. For example, the Protera.TM. line of visible
semiconductor lasers developed by Novalux, Inc. of Sunnyvale,
Calif. is based upon an extended cavity surface emitting laser
structure. An extended cavity laser designed to generate visible
light includes a number of optical elements to stabilize the
optical characteristics over a range of operating conditions during
the life of the laser. Additionally, a nonlinear crystal may be
included for frequency doubling a pump light source. The optical
elements must be initially aligned and kept in proper alignment,
which increases the cost and complexity of manufacturing. Generally
speaking, the manufacturing cost of high power visible
semiconductor lasers increases with each additional optical element
added to the packaged optical device. Moreover, each optical
element that requires a critical alignment adds a significant cost
to the final laser assembly.
[0012] Additionally, the form factor of a semiconductor laser is
also an important consideration in a projection display system.
There have been dramatic reductions over time in the total size of
projection display systems. See, or example, Derra et al. "UHP lamp
systems for projection display applications," J. Phys. D: Appl.
Phys. 38 (2005) 2995-3010, the contents of which are hereby
incorporated by reference. Miniaturization of the UHP lamp has
reduced the reflector size of the UHP lamp to less than 50.times.50
mm.sup.2 or less than about 2 inches on a side. UHP lamps with
reflectors having a diameter of 30 mm are also common, i.e., an
area corresponding to a square area (for design purposes) of about
one inch on a side. DLP chips are typically about 2 inches square
in size with an active (micro-mirror) region less than one inch
square (e.g., in the range of about 0.55'' to 0.75'' per side for
some DLP chips). Thus, UHP lamps are rapidly approaching small form
factors of about one to two cubic inches in size. For some
microdisplay applications even smaller volumes (e.g., one cubic
inch) are desirable. By way of comparison, the Protera.TM. line of
high power visible extended cavity surface emitting semiconductor
lasers developed by Novalux, Inc. of Sunnyvale Calif. generates 5
to 20 mW from an 11.6-cubic inch package having a length of about 4
inches (101.6 mm), a cross-sectional area of 1.79''.times.1.59''
(44.5.times.44.5 mm.sup.2) for each laser of a particular color.
The Protera.TM. package includes room for wavelength control
elements, such as etalons, polarization control elements, a surface
emitting gain element, a frequency doubling crystal, and other
control elements. However, in a projection display system a large
number of Protera.TM. lasers at different wavelength would be
required to have the range of wavelengths and total power required
such that the total volume of the set of Protera.TM. laser would be
immense compared to a conventional UHP lamp.
[0013] Another unresolved issue in the prior art is also how to
best utilize semiconductor lasers as part of a total light engine
solution. Semiconductor lasers have optical properties different
from those of UHP white light lamps. Thus, a simplistic direct
replacement of a UHP lamp with semiconductor lasers may not fully
exploit the potential benefits of semiconductor lasers in a
projection display system.
[0014] Therefore, in light of the above-described problems the
apparatus, system, and method of the present invention was
developed.
SUMMARY OF THE INVENTION
[0015] Arrays of vertical extended-cavity surface-emitting lasers
(VECSELs) include a gain chip having an array of emitters, an end
reflector, a nonlinear crystal for intra-cavity frequency
conversion, a wavelength control element, and a polarization
control element. The number of components that must be aligned
during packaging is reduced by forming an optical unit that
combines the functionality of at least two of the individual
components of the VECSEL.
[0016] In one embodiment, a volume Bragg grating acts as an optical
unit that has the functionality of both the end reflector and the
wavelength control element. Larger optical units may be formed by
attaching other optical components to the volume Bragg grating,
such as the nonlinear crystal or the polarization control
element.
[0017] In another embodiment at least one minor optical component
is attached to the gain chip. In one implementation an array of
lenses is attached to the gain chip. In still yet another
implementation a polarization control element is formed on the gain
chip.
BRIEF DESCRIPTION OF THE FIGURES
[0018] The invention is more fully appreciated in connection with
the following detailed description taken in conjunction with the
accompanying drawings, in which:
[0019] FIG. 1 is a perspective view illustrating a light source for
use in a projection display in accordance with one embodiment of
the present invention;
[0020] FIG. 2 illustrates an exemplary prior art extended cavity
surface emitting laser;
[0021] FIG. 3 illustrates an individual surface emitting diode
laser gain element for use in a light source in accordance with one
embodiment of the present invention;
[0022] FIG. 4 illustrates a projection display system in which
arrays of red, blue and green or more colors are focused into a
light guide that is subsequently imaged onto a micro-display light
valve and then re-imaged onto a screen in either a front or rear
projection display system;
[0023] FIG. 5 illustrates a projection display system in which a
diffractive optical element is used to convert the circular
Gaussian laser beam from each laser element in the light source
into a rectangular top-hat distribution that is subsequently
directed to efficiently cover the entire area of the light valve
and then subsequently imaged onto a front or rear projection
screen;
[0024] FIG. 6 illustrates a prior art projection system;
[0025] FIG. 7A illustrates a prior art projection system which
utilizes a color wheel;
[0026] FIG. 7B illustrates the beam footprint on the color wheel of
the prior art projection system of FIG. 7A during a blanking
period;
[0027] FIG. 7C illustrates the beam footprint on the color wheel of
a prior art projection system of FIG. 7A during an active
period;
[0028] FIG. 8 illustrates a prior art spatial modulator;
[0029] FIG. 9 illustrates a switching period in an element of the
spatial modulator of FIG. 8;
[0030] FIG. 10 illustrates a light engine system in accordance with
one embodiment of the present invention;
[0031] FIG. 11A illustrates calculated power changes in average
power in both fundamental (dashed curve) and second harmonic (solid
curve compared to the maximum achievable cw power at rollover as a
function of peak current for pulses that are comparable to or
longer than the thermal time constant;
[0032] FIG. 11B illustrates calculated power changes in average
power in both fundamental (dashed curve) and second harmonic (solid
curve compared to the maximum achievable cw power at rollover as a
function of peak current for pulses that are significantly shorter
than the thermal time constant;
[0033] FIG. 12A illustrates binary modulation of intensity for a
single pixel in an image from three colors by interleaving colors
at the sub-frame level in accordance with one embodiment of the
present invention;
[0034] FIG. 12B illustrates binary modulation of intensity for a
single pixel in an image from three colors by interleaving colors
at the frame level in accordance with one embodiment of the present
invention;
[0035] FIG. 13 illustrates a method of driving a single color laser
source with under-driving and over-driving during color-sequential
operation in accordance with one embodiment of the present
invention;
[0036] FIG. 14 illustrates an arrangement of components for
intracavity spectral filtering in accordance with the prior
art;
[0037] FIG. 15 illustrates a transmission response with wavelength
for a 300 micron thick etalon in accordance with the prior art;
[0038] FIG. 16 illustrates an extended cavity surface-emitting
laser with an intracavity interference filter for frequency
stabilization in accordance with one embodiment of the present
invention;
[0039] FIG. 17 illustrates a thin-film interference filter
structure for laser frequency stabilization in accordance with one
embodiment of the present invention;
[0040] FIG. 18 illustrates a sequence of thin-film layers for
forming an interference filter in accordance with one embodiment of
the present invention;
[0041] FIG. 19 illustrates theoretical transmission versus
wavelength for an interference filter in accordance with one
embodiment of the present invention;
[0042] FIG. 20 illustrates empirical transmission versus wavelength
for an interference filter in accordance with one embodiment of the
present invention;
[0043] FIG. 21 illustrates a surface-emitting extended cavity laser
array utilizing an interference filter for frequency control in
accordance with one embodiment of the present invention;
[0044] FIG. 22 illustrates empirical reflectivity versus wavelength
for a volume Bragg grating in accordance with one embodiment of the
present invention;
[0045] FIG. 23 illustrates an extended cavity laser utilizing a
volume Bragg grating in accordance with one embodiment of the
present invention;
[0046] FIG. 24 illustrates an extended cavity laser array utilizing
a volume Bragg grating in accordance with one embodiment of the
present invention;
[0047] FIG. 25 illustrates functional block elements of a vertical
extended cavity surface emitting laser with intracavity frequency
doubling in accordance with one embodiment of the present
invention;
[0048] FIG. 26 illustrates the functional blocks of FIG. 25 with
the functionality of an end mirror and wavelength control element
combined in one optical unit to reduce alignment complexity in
accordance with one embodiment of the present invention;
[0049] FIG. 27 illustrates the functional blocks of FIG. 25 with
the functionality of an end mirror, wavelength control element, and
nonlinear crystal combined in one optical unit to reduce alignment
complexity in accordance with one embodiment of the present
invention;
[0050] FIG. 28 illustrates the functional blocks of FIG. 25 with
the functionality of an end mirror, wavelength control element,
nonlinear crystal, and polarization control element combined in one
optical unit to reduce alignment complexity in accordance with one
embodiment of the present invention;
[0051] FIG. 29 illustrates the functional blocks of FIG. 25 with
the functionality of minor optical components integrated into other
units to reduce alignment complexity in accordance with one
embodiment of the present invention;
[0052] FIG. 30 illustrates an extended cavity surface emitting
array utilizing a volume Bragg grating and turning mirror in
accordance with one embodiment of the present invention;
[0053] FIG. 31 illustrates an extended cavity surface emitting
array having an optical unit including a volume Bragg grating,
nonlinear crystal, and waveplate in accordance with one embodiment
of the present invention;
[0054] FIG. 32A is a perspective view illustrating a packaging
design for an array of surface emitting lasers of one color in
accordance with one embodiment of the present invention;
[0055] FIG. 32B illustrates a cross-sectional view of the package
of FIG. 32A;
[0056] FIG. 32C illustrates a perspective view of three packaged
arrays of lasers in accordance with one embodiment of the present
invention;
[0057] FIG. 33 illustrates a diagram of frequency doubled surface
emitting laser diode with electrical drive current restricted by
proton implantation;
[0058] FIG. 34 illustrates the effect of lateral optical pumping
for a large diameter VECSEL device; and
[0059] FIG. 35 is a chart comparing the performance of surface
emitting laser arrays with conventional LED and UHP sources as
light sources for projection display.
[0060] Like reference numerals refer to corresponding parts
throughout the several views of the drawings.
DETAILED DESCRIPTION OF THE INVENTION
I. Basic Architecture Of Extended Cavity Surface Emitting Light
Source For Projection Displays
[0061] FIG. 1 is a profile view showing a light source 100 for
generating light at several different colors required by a light
processing (LP) system. In a red-green-blue (RGB) LP system the
light source produces red, green, and blue light. A first array 105
of semiconductor lasers is used to generate a plurality of beams
107 of blue light from two or more individual lasers. A second
array 110 of semiconductor lasers is used to generate a plurality
of beams 113 of red light from two or more individual lasers. A
third array of semiconductor lasers 115 is used to generate a
plurality of beams 118 of green light from two or more individual
lasers. Thus, light source 100 includes different sets of lasers.
An individual set of two or more lasers generates a particular
color of light used in the LP system. However, as described below
in more detail, in a preferred embodiment individual lasers in a
set are designed to be substantially incoherent (e.g., un-phased)
with respect to other lasers in the same set in order to reduce
speckle. In FIG. 1 each laser array 105, 110, and 115 is
illustrated as generating four beams from four individual laser
emitters (not shown). However, more generally each laser array 105,
110, 115 may have an arbitrary number of individual lasers.
[0062] Note that in one embodiment individual lasers and optical
elements (not shown) in light source 100 are preferably arranged
such that the individual beams 107, 113, and 118 of arrays 105,
110, and 115 at least partially overlap such that light source 100
provides uniform illumination of a spatial light modulator (not
shown). The spatial light modulator may be a light valve, such as a
digital-mirror light valve (sometimes known as a "digital light
valve") having micro-mirrors, a light valve formed from an array of
liquid crystal elements, or a liquid crystal on silicon (LCOS)
spatial modulator. The initial overlap of individual beams 107,
113, and 118 may, for example, be only partial and increased by
additional optical elements (not shown).
[0063] One or more of the arrays 105, 110, and 115 may utilize a
nonlinear frequency converter 120, 125, or 130 to convert a source
emission frequency into an output color of a different frequency.
Nonlinear frequency conversion permits light emitted at one source
frequency to be converted into another frequency. For example, in
one embodiment the red, blue and green wavelengths are generated by
frequency doubling a source laser frequency generated by a
semiconductor gain element in individual lasers. In another
embodiment, the red lasers operate directly in the red wavelength
range but the green and blue colors are generated by frequency
doubling. Alternatively, the red lasers can be made up of an array
of green lasers that are parametrically down-converted by a
nonlinear optical process. The nonlinear frequency converters may
be disposed within a laser resonator of each individual laser or be
disposed outside of the resonator. Examples of nonlinear crystals
that may be used as nonlinear frequency converters include
periodically poled lithium niobate, periodically poled lithium
tantalate, and periodically poled KTP. These nonlinear crystals may
use appropriate dopants such as magnesium oxide to improve their
reliability and/or poling quality.
[0064] Output couplers 135, 140, and 145 may be included to couple
the light and to provide additional frequency stability. For
example, the output couplers 135, 140, and 145 may include volume
Bragg gratings for coupling light and providing frequency control
of individual sub-arrays. In one embodiment one or more of the
output couplers comprises volume Bragg gratings, for example.
Additionally, one or more wavelength controlling elements (not
shown) may be included to control the wavelength of each laser.
Examples of wavelength controlling elements include intra-cavity
surface gratings, volume gratings, solid etalons, thin-film-coating
etalons, and Lyot filters. These wavelength-controlling elements
may be separate or may be monolithically combined with other
intracavity elements for compactness and cost reduction
reasons.
[0065] One aspect of the present invention is that the total
optical power scales approximately with the total number of lasers
in light source 100. As an illustrative example, light source 100
may be formed from three different semiconductor dies, each having
at least one row of lasers. Additionally, power output may be
further increased by including additional die in a tile
configuration, such as two or more dies for at least one the arrays
105, 110, or 115. The packaged light source 100 may, for example,
form an array of emitters having a total area on the order of one
square centimeter. In one embodiment (not shown) the output from
two or more light sources 100 is optically combined. Additionally
optical techniques may also be used to combine the output of
individual arrays 105, 110, and 115 with arrays emitting the same
color. Examples of techniques to optically combine the light output
of two or more arrays 105, 110, and 115 or two or more light
sources 100 include dichroic optics or polarization beam combining
optics.
[0066] Another aspect of the present invention is that one or more
of the red, green, or blue arrays 105, 110, or 115 may be formed
from extended cavity surface-emitting semiconductor lasers. FIG. 2
illustrates an exemplary prior art extended cavity surface emitting
laser 200. Extended cavity surface-emitting semiconductor lasers
are a class of semiconductor lasers that have a number of
advantages over edge emitting semiconductor lasers or conventional
surface emitting lasers. Extended cavity surface emitting lasers
typically include at least one reflector disposed within a
semiconductor gain element. For example, an intra-cavity pair of
Bragg mirrors 205 grown on either side of a quantum well gain
region 210 form a Fabry-Perot resonator to define the operating
wavelength of the fundamental laser wavelength. An additional
external reflector 215 spaced apart from the semiconductor gain
element defines an extended cavity of an optical resonator,
providing additional wavelength control. By appropriate selection
of the quantum well gain region 210, Bragg mirrors 205, and
external reflector 215 a fundamental wavelength can be selected
within a large range of wavelengths. The fundamental wavelength, in
turn, may then be frequency doubled by frequency doubling optical
crystal 220 to generate light at a desired color.
[0067] Additional optical elements, including optical frequency
doubling optical crystal 220 or other frequency conversion devices,
may be included in the optical resonator of the extended cavity for
intra-cavity frequency conversion. As another example, a saturable
absorber (not shown) may be included in the extended cavity (or
within the semiconductor gain element) to form a mode-locked laser.
An efficient mode-locked extended cavity semiconductor laser is
described in copending U.S. patent application Ser. No. 11/194,141,
"Apparatus, System, and Method for Wavelength Conversion of
Mode-Locked Extended Cavity Surface Emitting Semiconductor Lasers,"
filed on the same day as the present application, the contents of
which are hereby incorporated by reference.
[0068] Extended cavity surface-emitting semiconductor lasers
developed by the Novalux, Inc. of Sunnyvale, Calif. have
demonstrated high optical power output, long operating lifetimes,
accurate control of laser wavelength, control of spatial optical
mode, provide the benefit of surface emission for convenient
manufacturing and testing, and may be adapted to include optical
frequency conversion elements, such as second harmonic frequency
doublers, to generate light at the red, green, and blue colors.
Additionally, an array of high power extended cavity surface
emitting lasers may be fabricated on a single die such that light
source 100 may be fabricated in a highly manufacturable process
that utilizes only a small number of different semiconductor dies,
frequency converters, and frequency control elements. As a result,
a light source 100 utilizing extended cavity surface emitting
arrays may be manufactured that has a useful power output at red,
green, and blue colors that is greater than conventional white
light LP systems and which may be manufactured at a reasonable
price due to the high manufacturability. Background information
describing individual extended cavity surface emitting
semiconductor lasers and frequency-doubled surface emitting lasers
developed by the Novalux, Inc. are described in U.S. Pat. Nos.
6,243,407, 6,404,797, 6,614,827, 6,778,582, and 6,898,225, the
contents of each of which are hereby incorporated by reference.
Other details of extended cavity surface emitting lasers are
described in U.S. Pat. App. Ser. Nos. 10/745,342 and 10/734,553,
the contents of which are hereby incorporated by reference.
Additionally details regarding exemplary extended cavity
surface-emitting semiconductor lasers are described below in more
detail.
[0069] In one embodiment the lasers in light source 100 are
designed to be operated in a pulsed mode. Electrical connectors
(not shown) are provided to drive each individual laser of each
array 105, 110, and 115. A laser controller 180 is used to regulate
drive current and/or voltage of each laser array 105, 110, and 115.
Laser controller 180 preferably permits individual laser arrays
105, 110, 115 to be operated in a pulsed mode. In one embodiment,
laser controller 180 permits individual laser arrays 105, 110, and
115 to be operated sequentially, i.e., to generate red light during
one time interval, green light during another time interval, and
blue light during another time interval, thereby eliminating the
necessity in a LP system for a color wheel to filter light. In one
embodiment individual lasers and subgroups of lasers are
independently addressable by laser controller 180. Laser controller
180 may also use feedback from optical monitors (not shown) to
adjust drive current/voltage to maintain selected output
characteristics. For example, laser controller 180 may adjust drive
current or pulse width to maintain a desired grey scale in a LP
system. In one embodiment a manufactured laser array 105, 110, and
115 includes at least one redundant laser. In this embodiment,
laser controller 180 determines an initial set of operable lasers
that are used in each array. The redundant lasers kept in reserve
are then turned on as required to replace lasers that have failed,
thereby increasing the useful lifetime of light source 100.
[0070] In one embodiment, each laser array 105, 110, and 115 is
driven in a pulsed mode. A pulsed mode operation provides a number
of benefits, including reduced heat dissipation. Additionally a
pulsed mode of operation may also, depending upon pulse parameters,
provide beneficial spectral broadening. Light valves typically have
a characteristic time that is required for them to change
reflective or transmissive states and an associated minimum
on-time. In a pulsed mode of operation light source 100 preferably
generates pulses having a high repetition rate such that the period
between optical pulses is small compared with the minimum on-time
of the light valve. Consequently, a high repetition rate permits a
pre-selected minimum number of laser pulses to strike a
micro-mirror of a digital light valve during its shortest possible
on-cycle, improving grey scale resolution. An exemplary repetition
rate is as fast as 1 MHz, which is considerably faster than
conventional frame repetition rates used in television, video, and
computer graphics. More generally the repetition rate can be
selected based upon the characteristics of particular LP systems.
For example, the laser repetition rates are preferably high enough
to be compatible with achieving a full grey scale with commercially
available digital mirror light valves or liquid crystal light
valves that are used in boardroom projectors and home
television.
[0071] Light source 100 is preferably designed to substantially
reduce speckle in a LP system to acceptable levels, i.e., to reduce
speckle to a level that provides people viewing an image in a LP
system with an acceptable viewing experience. As is well known in
the field of optics, speckle results when coherent light is
reflected from scatter centers. The resulting interference
generates bright and dark patches having a speckled appearance.
Speckle can be numerically characterized by a peak-to-valley
intensity modulation of a speckle pattern, i.e., the percentage
variation in peak-to-valley intensity between bright and dark
patches. One rule that is sometimes used in the LP art is that a
speckle modulation of about 4% or less is required in a LP system
to achieve an acceptable viewing experience.
[0072] In the context of LP systems illuminated by a laser, speckle
occurs as coherent light reflects from scattering centers in the
light valve or in other optical elements of the LP system. Speckle
generates optical non-uniformities that degrade the optical quality
of the projected image. Speckle is a particularly severe problem in
LP systems receiving highly coherent light from a single laser
source. As is well known in the field of optics, coherence is
associated with the ability of light waves to constructively and
destructively interfere. A single laser source that has a narrow
distribution of optical frequencies, phase, and spatial attributes
is highly coherent and thus is prone to generating speckle.
[0073] In accordance with one embodiment of the present invention,
speckle is reduced in a LP system using light source 100 by several
different techniques that may be used alone or in combination to
reduce coherence in light source 100 by increasing the distribution
of optical attributes (e.g., phase, spectral width) of light source
100. First, increasing the number of lasers of the same color in
light source 100 that are substantially incoherent with respect to
each other will tend to reduce speckle. Even though each laser in a
single array 105, 110, or 115 emits coherent laser light at the
same color, un-phased lasers of the same color having different
phases are incoherent to each other in regards to the optical
interference effects that cause speckle. As a result, the
speckle-induced intensity variations for each un-phased laser is
independent of that of other lasers such that the combined
illumination uniformity for an array 105, 110, or 115 of un-phased
lasers improves as the number of un-phased lasers increases. In
particular, speckle tends to decrease inversely with the square
root of the number of un-phased lasers of the same color having
equal amplitude. The lasers in each array are thus preferably
designed to operate un-phased with respect to each other, i.e., an
individual array 105, 110, or 115 is designed to not provide
significant inter-laser feedback that would significantly lock the
frequency and phase of two or more lasers in the array to each
other. Second, since spectral broadening reduces speckle, any
operating mode that broadens the spectra will tend to decrease
speckle. Examples of techniques to broaden the spectra of
semiconductor lasers include operating lasers in a pulsed mode with
the pulse parameters (e.g., on time) selected to generate spectral
broadening; high frequency modulation of lasers selected to
generate spectral broadening; mode-locking lasers to generate
spectral broadening, and sub-combinations thereof such as operating
mode-locked lasers in a pulsed mode. Third, the individual lasers
in an array 105, 110, and 115 may be designed to operate with
multiple frequency, axial, or spatial modes to increase the
frequency, phase, and directional (angular) distribution of each
laser. Fourth, an optical element may be used to scramble the
direction, phase, and polarization information of the beams 107,
113, and 118 to reduce coherence.
[0074] It is desirable that the extended cavity lasers used in
laser arrays 105, 110, and 115 have stable optical characteristics.
Additionally, if frequency doubling is used it is desirable to
minimize dephasing effects that reduce the efficiency of the second
harmonic conversion process. FIG. 3 illustrates an extended cavity
surface emitting laser 300 having a lens 310 for stabilizing the
optical mode and reducing deleterious dephasing in a frequency
doubling process. Each laser gain element 305 includes an optical
gain region and may also include one or more distributed Bragg
reflectors (not shown). In one embodiment an etched lens 310 is
formed on the semiconductor substrate 315. In an alternate
embodiment an external lens (not shown) is used instead of etched
lens 310 with the external lens formed using conventional optical
materials, such as optical glass. A profile of an exemplary
fabricated etched lens 310 is illustrated in detailed inset 350. A
nonlinear optical material 320 is contained in the laser resonator.
A resonant mirror 325 is deposited directly on the nonlinear
material or may be a grating device, both of which are designed to
control the output wavelength to match the phase-matching
wavelength of the nonlinear optical material 320. However, it will
be understood that other techniques may be used to fabricate
individual elements of arrays.
[0075] The intra-cavity lens 310 is used to provide a stable
transverse mode (preferably a TEM.sub.00 mode) in a semi-confocal
configuration. In this arrangement, the infrared light (i.e., the
fundamental frequency light generated by the semiconductor which is
to be frequency doubled) is focused into the nonlinear optical
material 320 while the forward going harmonic light is transmitted
through a flat output coupler (not shown) and the backward harmonic
light is reflected from the surface of the lens 310 that is coated
to be highly transparent to the fundamental wavelength and highly
reflective at the harmonic wavelength. This lens element acts as a
convex mirror for the harmonic light and will have a reduced
intensity in the nonlinear material as it overlaps the fundamental
wavelength laser beam.
[0076] The reduced intensity in the overlap region with the
fundamental laser beam in the nonlinear optical material 320 will
have the advantage of reducing any dephasing effect that would
reduce the overall nonlinear conversion efficiency. The beam
divergence of the forward and backward waves is about the same. The
mode waist position of the harmonic light of the forward wave is
near the output mirror position while that for the backward wave is
at the virtual image of the reflective mirror. These two beams will
propagate in the same direction but with somewhat different beam
angles.
[0077] The focusing lens 310 in the cavity can be made by etching
the surface of the semiconductor or by the introduction of a glass
or binary optical lens, all of which can be made in an array
format. The introduction of such a lens in the cavity would reduce
the variable effects of thermal lensing usually present in most
laser systems that have materials with a positive dn/dt (change of
refractive index with temperature). However, an embodiment that
relies on the thermal lens only to stabilize the spatial mode of
the cavity is also within the scope of this invention. In this
case, a more conventional method for extracting the
backward-propagating second-harmonic beam can be used, for example,
via a 45-degree beamsplitter.
[0078] A nonlinear crystal 320 is placed either at or near the
output mirror (not shown) or it may have a mirror 325 deposited
directly to it that is highly transmissive at the harmonic and
highly reflective at the fundamental wavelength. If a separate
output mirror is used it may also be comprised of a volume
holographic grating for wavelength selection or any other
wavelength resonant reflector such as a replica grating or a
resonant etalon reflector. Alternatively, a mirror may be a
conventional glass element coated with a broadband coating and a
wavelength selector such as a solid or thin-film deposited etalon
may be placed into the cavity at a tilt angle. Such an element
(grating or etalon) may be separate or integrated with other
elements (for compactness and cost reduction). For example, a
conventional grating or a thin-film interference filter may be
deposited on a wedged surface of the nonlinear crystal or the
second (intracavity-looking) surface of the output mirror. The
output mirror may also be a convex reflector designed to provide an
intermediate focus point in the cavity where the nonlinear optical
material 320 is placed, thereby doubling the effective path length
in the nonlinear crystal.
[0079] It will be understood that light source 100 may be adapted
for use with different types of spatial modulators besides light
valves. In one embodiment, light source 100 may also have an
optical lens system designed to generate a single beam for use in a
scanning optical system. In a scanning optical system the spatial
modulator comprises mirrors that are used to scan a laser beam in
an X-Y raster motion. For a scanning optical system, each of the
beams in the array should preferably be nearly diffraction
limited.
[0080] FIG. 4 is a block diagram of a LP system 400 in accordance
with one embodiment of the present invention illustrating a
schematic layout of the optical path from a light source 100 to a
projection screen 405. The beams from light source 100 can be
focused into a light guide 410, rectangular in most cases, and then
focused onto a light valve 420 (e.g., a digital light valve) after
the beam emerges from light guide 410. The beams generated by light
source 100 will have some initial overlap. However, the light guide
may be selected to increase the overlap of the beams. In
particular, internal reflection, diffraction, mode conversion, or
other electromagnetic properties of a waveguide may be used to
scramble the beams within light guide 410. In this case, all the
beams will overlap with some degree of uniformity with reduced
speckle that might be present in each of the beams separately. In
the case of mode-locked laser arrays, additional reduction of
speckle will occur from the spectral broadening of the mode-locked
pulses. Pulsed devices that have spectral broadening due to
chirping or operation in several spatial or spectral modes will
also have some degree of speckle reduction.
[0081] The individual lasers in system 400 may be un-phased. Pulsed
operation will tend to broaden the spectrum of the individual
lasers, therefore, reducing the degree of their coherence and
reducing speckle. A single lens 430 will focus all of the lasers
beams that are traveling parallel to one another to one focal spot
with a spot size determined by the diameter of each beam and the
focal length of lens 430. Each sub-aperture may be filled by
expanding the laser beam to the pitch of the array and then
subsequently focused with a single lens 430 to provide the smallest
spot size for all the beams. The input laser beams undergo several
bounces within light guide 410 that tend to scramble the beams so
that the output light exits light guide 410 with a rectangular
pattern that is uniformly filled to match the digital light
valve.
[0082] In one embodiment system 400 includes a laser controller 180
for controlling the laser array and a light valve controller 480
for controlling the light valve. The two controllers preferably
have a link 490 for communicating with each other. In one
embodiment, laser controller 180 controls operational attributes of
light source 100 such as repetition rate. A sensor (not shown) may
be included for the laser controller 180 to monitor laser output
and control the output power. For the case of a mode-locked laser,
the laser controller 180 can also control the operation of
saturable absorbers used to mode-lock the lasers.
[0083] Laser controller 180 also determines whether the red, green,
or blue lasers of light source 100 emit light during a particular
portion of time during a frame. In one embodiment, LP system 400
does not have a color filter wheel. Instead, the red, green, and
blue portions of a frame are generated by synchronizing the red,
green, and blue lasers of light source 100 to operate in a burst
mode during different times of the frame. In a burst mode, only one
set of lasers (e.g., red, green, or blue) is used to generate a
train of optical pulses. Thus, in each frame to be displayed, the
laser controller and light valve controller 480 communicate with
each other to synchronize turning on only the red lasers when the
red portion of pixels are to be projected by light valve 420, turn
on only the green lasers when the green portion of the pixels are
to be projected by the light valve 420, and turn on only the blue
lasers when the blue portions of the pixels are to be projected by
light valve 420. This coordination allows a system with a single
digital light valve chip to eliminate a color wheel filter,
improves reliability and efficiency, and improves the quality of
the displayed image.
[0084] Additionally, in some embodiments the laser repetition rate
is adjusted by laser controller 180 to account for the response
time of individual elements in light valve 420, such as individual
mirrors in a digital light valve. Digital light valves control grey
scale by pulse width modulation of the on/off response of the
micro-mirrors. An individual micro-mirror of a digital light valve
has a rise and fall time associated with rotating the micro-mirror
between an off-state to an on-state. A burst of laser light of a
specific color (e.g., red) is a sequence of pulses. As previously
mentioned, selection of a comparatively high nominal pulse rate
facilitates control of gray scale. A high pulse rate permits a
comparatively small minimum mirror on-time to be used (since a
number of optical pulses in the burst will be capable of being
deflected to the projection screen).
[0085] In one embodiment the pulse repetition rate of each laser in
light source 100 is adjusted to optimize the optical response in
the most critical rise/fall portions of the micro-mirror of a
digital light valve. For example, the mirror of a digital light
valve tilts within a finite range of motion, such as 10 to 15
degrees. However the last 1 to 2 degrees of tilt have a great
impact on control of the grey scale. By adjusting the repetition
rate based on the turn on/turn off response of the micro-mirror,
the optical response of the digital light valve may be optimized.
It will also be understood that with sufficient synchronization of
the digital light valve and each laser array 105, 110 and 115 that
the turn-on and turn-off of each laser array burst may be
advantageously timed to initiate or stop a train of pulses at
precise times with respect to the turn on or turn off phases of the
micro-mirrors. Additionally, it will be understood that the
repetition rate of a laser array 105, 110, and 115 may be
dynamically varied during a burst.
[0086] In one embodiment, the effective pulse rate is increased by
electrically dividing each array 105, 110, 115 into separate
sub-sections ("sub-arrays") that are driven with different phases
to increase the effective pulse rate. In this embodiment the
different sub-arrays of a single color are each pulsed with a
separate driver delayed in phase with respect to each other. The
phase difference may, for example, be selected to be a fraction of
the pulse repetition rate divided by the number of sub-arrays in
order to increase the total repetition rate of pulses of each
color. This would overcome any pulse repetition rate limitations
for any individual element of the array in order to achieve any
appropriate gray scale and to be compatible with the digital mirror
pulse widths and rise and fall times of the micro-mirrors.
[0087] FIG. 5 illustrates an embodiment of a projection system that
includes an optical lens system 505 (illustrated by the
checkerboard pattern) designed to convert the optical output from
each laser in light source 500 to a profile that matches that of a
light valve 540. For example, an individual laser element 510 has
an optical element that converts its output into an optical output
that matches that of the light valve 540. Similarly, another
individual laser element 520 has an optical element that converts
its output into an optical output that matches that of the light
valve. A benefit of this arrangement is that the optical uniformity
on the surface of light valve 540 is independent of which
individual laser elements in light source 500 are operative. That
is, the result of optical lens system 505 is that the output from
each laser overlaps at light valve 540 with the same profile. Thus,
failure of one or more of the individual laser elements 510 or 520
will not affect the illumination uniformity across light valve 540.
Consequently, in the event that an individual laser element fails
the power output of other laser elements can be increased to
compensate for the loss in power with no change in illumination
uniformity. Moreover, this arrangement facilitates the use of
redundant lasers that are turned on as required to maintain the
power output.
[0088] In one embodiment optical lens system 505 projects each
laser beam in the array of light source 500 into a rectangular
top-hat intensity profile 510 that is imaged onto the digital light
valve 520. An optical element, such as a diffractive optical
element, is used to convert the circular Gaussian laser beam from
each element in light source 500 into a rectangular top-hat optical
distribution that is subsequently directed to efficiently cover the
entire area of the light valve and then subsequently imaged onto a
front or rear projection screen. This lens system may be
reflective, diffractive, or transmissive and may be made from a
glass lens array or a digital optical lens system, for example.
Digital optical lens systems may be made from a variety of
materials, such as plastics. Digital optical lens systems are
available from a variety of vendors, such as MEMS Optical, Inc, of
Huntsville, Ala. As one example, optical lens system 505 may
comprise an assymetric binary optical lens arrangement designed to
generate a top-hat intensity distribution for an array of lasers
that each generate a circular Gaussian laser beam. More generally,
optical lens system may be designed using optical modeling
techniques taking into account a desired intensity profile to be
imaged onto a light valve 540, selecting an arrangement of lasers
in light source 500, and the beam properties of each laser in light
source 500.
[0089] The optical lens system 505 permits each laser in the array
to have an optical beam profile that approximately matches the
profile of the digital light valve. The optical lens system is
preferably designed so that each laser in the array has a nearly
identical rectangular top-hat intensity profile projected onto the
digital light valve. This allows a near perfect match of each laser
to the digital light valve. Moreover, the total optical system of a
LP system may be less costly in spite of the addition of optical
lens system 505 because a conventional three prism dichroic lens
system can be eliminated in a LP system. Additionally, the failure
of an individual laser can be accounted for by adjusting the power
of other lasers. The system of FIG. 5 allows the projection system
to have a uniform light intensity and to maintain this intensity
over the life of the system by being able to control the total
optical power from the array. This may be accomplished by
increasing the drive current to the array or by having additional
(redundant) laser elements in the array that may be turned on at a
later time.
[0090] Each surface emitting laser array 105, 110, and 115 may be
wired so that all elements can be in series or some combination of
series and parallel operation. This would allow for a power supply
with improved efficiency due to reduced currents and reduced
resistive heating. In addition, the series wired array would be
driven with one current to avoid thermal runaway. Further, each
color sub-array can be driven for only one third of the time in
sequence in contrast to lamps which are on all of the time and use
a color filter wheel to select only one color while rejecting the
other wavelengths from the lamp.
[0091] While the present invention has been described in regards to
a light source generating three primary colors, it will also be
understood that light sources 100 and 500 may be adapted to
generate a different sequence of colors or more than three colors
by selecting the characteristics and numbers of laser arrays to
achieve beams with the desired output colors. Thus it will be
understood that the present invention is not limited to use in
light projection systems based on red, green, and blue (RGB) light
sources. For example, if a light projection system requires a
different set of colors than RGB the light sources 100 and 500 of
the present may be adapted to include arrays of lasers generating
each of the different colors required by the light projection
system. It will also be understood that nonlinear frequency
conversion may be extended to frequency tripling or
quadrupling.
[0092] Some types of spatial modulators work most efficiently with
polarized light. As one example, certain types of liquid-crystal
modulators modulate polarized light. It will be understood that in
some embodiments light sources 100 and 500 are designed with the
polarization selected to facilitate modulation by a spatial
modulator that modulates polarized light.
II. Methods for Driving Laser Sources for Display Applications
[0093] As previously described, semiconductor lasers have a number
of advantages for displays, including high-brightness, low etendue,
extended color gamut, and the capability for modulation. This
latter advantage is particularly important for displays due to the
inherent modulation required by motion video. Careful use of
modulation in the laser sources can lead to benefits in the overall
display system.
[0094] Furthermore, semiconductor lasers sources for displays may
utilize non-linear optical conversion to create the desired display
wavelengths. All of these processes scale in a supra-linear way
with power. Consequently, pulsing frequency-doubled semiconductor
lasers can lead to increased output or efficiency on an average
basis, even when the average input power is unchanged or reduced.
However, in order to fully exploit the advantages of
pulse-modulating semiconductor laser sources requires coordinating
the drive parameters of the lasers with the spatial light
modulator.
[0095] Some of the issues associated with replacing a conventional
white light source with semiconductor lasers may be understood with
regards to FIG. 6, which illustrates a conventional projection
display system consisting of five components: a white light source,
a color filter, a spatial modulator, a projection lens assembly,
and a projection surface. In some cases one or more of the
components may be combined, such as when the display system
utilizes a cathode-ray tube as both a light-source and a spatial
modulator, or when the light source produces a set of discrete
colors, and so effectively contains a color filter.
[0096] When projection systems are used to display motion video,
they do so by generating a series of static images at a rate faster
than the human eye can follow. This rate, commonly referred to as
the fusion frequency, is typically 24-30 Hz. Video signals for
television are typically encoded at 60 interlaced images per
second. Interlacing is a method of speeding up the image display
rate by expressing images as a series of stripes or lines, and
displaying every other line per image in an alternating fashion. As
only half the lines are drawn in any one image, the rate can be
doubled. A single image in such a display system is commonly
referred to as a frame. In many projection systems, the image is
formed by the combination of at least three color channels,
typically red, green, and blue. In some cases, the image for each
color is generated in a sequential fashion by a single modulator.
This is referred to as color sequential operation.
[0097] FIGS. 7A, 7B, and 7C illustrates in more detail how a
conventional projection system utilizes a color wheel for color
sequential operation. FIG. 7A illustrates a conventional projection
system having a color wheel. The white light source is typically an
arc lamp with condensing optics, the color filter is formed as a
series of filters disposed about a rotating axis, commonly referred
to as a color wheel, such that as the assembly rotates, the light
source is filtered into component colors in a sequential fashion
such that the colors occur in a sequence (e.g.,
red-to-green-to-blue as the color wheel rotates). The maximum
length of time that each color in the sequence lasts depends upon
the rotation speed of the color wheel and the number of segments in
the wheel. The spatial modulator in this example is comprised of a
single modulator that acts upon each color in sequence to form the
combined image on the screen. All of this happens during each frame
of the video signal. One consequence of using such a filter is that
as one of the boundaries between the filters is swept across the
light sources light of mixed colors is created. This light must be
prevented from reaching the projection surface. This is typically
accomplished by setting the spatial modulator to a minimum
transmission state. Hence these periods are commonly referred to as
blanking periods. FIG. 7B, illustrates the beam footprint on the
color wheel during a blanking period. FIG. 7C illustrates the beam
footprint on the color wheel during an active period when light of
a single color illuminates the display modulator. Thus, as an
example, during the rotation of the color wheel an exemplary
sequence over time might be red light for a first time period, a
blanking period, green light for a second time period, a blanking
period, then blue light for a third time period, and so on. As an
illustrative example, in many DLP systems the color wheel rotates
at about 120 Hz to support interlacing frames generated at a rate
of 60 Hz. In some cases, the color wheel is further divided into
six segments to generate an equivalent red-green-blue rotation rate
of 240 Hz to support high resolution displays and to reduce visual
artifacts. Thus, in this example the colors will change in sequence
such that a complete R-G-B sequence is repeated at a rate of
120-240 Hz to generate complete color frames at 60 Hz. Note that
even if the effective rotation rate of the color wheel can be
increased somewhat by increasing the rotation rate or the number of
segments the blanking periods impose a practical limit on the
effective rotation rate due to the finite width of the blanking
periods.
[0098] An alternate approach in the prior art (not shown) is to
replace the color wheel with a filter system that spatially
separates the component colors, so that each color impinges upon a
separate modulator. The modulated colors are then recombined before
being projected on the screen. In this system, the component colors
are handled in parallel eliminating some of the problems associated
with color sequential operation. The extra modulators and optics
required make this parallel arrangement of spatial modulators more
expensive than a color sequential system. Thus, a color sequential
system is commonly used in consumer products. However, the faster
modulators required by a color sequential system can offset the
other cost savings, which means that systems based on slower
modulators, e.g., high temperature, poly-silicon, liquid-crystal
based modulators, enjoy significant market share.
[0099] A class of projection systems that is of particular interest
is one which utilizes an array of fast, binary-type spatial
modulators, i.e., the modulator has only two states, on and off,
for each pixel. Referring to FIG. 8, a spatial modulator 800 has a
number of pixel elements 805, such as micro-mirrors, that may be
turned either or off during a particular switching period. As
illustrated in FIG. 9, the light intensity for a particular pixel
may be selected by determining the number of switching periods
during a frame that a corresponding pixel element 805 in spatial
modulator 800 is in an on-state and illuminated by the color wheel
with light of a particular single color. In conventional systems
the color wheel illuminates the spatial modulator for a large
number of switching periods with light of one color before the
wheel rotates into position for the next color. Thus, for example,
the red color component of an individual pixel depends of upon the
number of switching periods during a frame that a corresponding
pixel element 805 in the spatial modulator is in an on-state and
receiving red light via the color wheel.
[0100] An example of such a digital light modulator is the DLP.TM.
modulator from Texas Instruments. These devices achieve modulation
depth by breaking a single frame into a series of sub-frames,
typically 2, where n is typically between 6 and 10. To achieve a
particular light level, the modulator changes the number of
sub-frames that a particular pixel is in the ON state during the
frame. The color with which a particular pixel of a frame is
displayed depends upon the number of sub-frames that the
corresponding element of the modulator is in an "on" state for each
color. Individual pixels are turned on for some fraction of the
sub-frames, depending upon the brightness required at a given
pixel. Ideally the on and off sub-frames are interleaved during the
formation of a frame. Note that in many commercial system that the
sub-frame rate is much higher than the effective rotation rate of
the color wheel.
[0101] One drawback of conventional color sequential operation is
that if the light path is interrupted during the formation of a
single frame, for instance by motion of the viewer's head or eye,
or by physical occlusion of a portion of the display, the frame
appears to be strongly biased toward one or two colors. For
example, if the light path is partially or totally interrupted when
the red filter is aligned, the frame will be biased towards green
and blue by virtue of the reduction in intensity of red light.
Additionally, if only a portion of the light path is blocked then
the frame will have a color bias across the frame. The color
biasing associated with an interruption of the light path can
result in a negative viewer experience in a conventional projection
display system.
[0102] Referring to FIG. 10 accordance with the present invention,
a conventional white light source is replaced with a light engine
assembly 1000 that includes a set of lasers 1005 of different
colors, such as red, green, and blue lasers. The light engine
assembly 1000 optically couples the set of lasers 1005 to a spatial
modulator 1010. Control electronics 1020 includes a laser
controller 10255 laser drivers 1030, and a spatial modulator
controller 1035. The set of lasers 1005 may, for example comprises
sets of any of the laser arrays described in the present patent
application.
[0103] Each color laser of the set of lasers 1005 is separately
driveable in a pulsed mode. By using pulsed lasers for the light
source, a color wheel is not required. Additionally, the pulse rate
can be as fast or faster than an individual switching period
(sub-frame) of the spatial modulator 1010. The control electronics
1020 synchronizes the laser pulses generated by the pulsed lasers
with the switching periods of the spatial modulator, so that an
whole number of laser pulses of one color are emitted during a
single sub-frame. When the pulse rate of the laser source is fast
compared to typical application timescales or thermal timescales in
the laser, e.g. >1 kHz, then the laser source is commonly
referred to as quasi-continuous wave (QCW). QCW semiconductor
lasers have peak powers that are much greater than their average
powers. This is due, in part, to the fact that operating a laser in
a pulsed mode limits heating effects which decrease the output of
semiconductor lasers at a fundamental frequency. Additionally, for
lasers that include frequency-doubling elements, the conversion
efficiency also increases nonlinearly with respect to the pump
laser power. Consequently pulsed mode operation of frequency
doubled semiconductor lasers is particularly beneficial to increase
output power. Moreover, pulsed mode operation also produces
spectral broadening, reducing speckle.
[0104] FIGS. 11A and 11B illustrate theoretical and experimental
investigations by the inventors of the present patent application
for extended cavity lasers with intracavity frequency doubling
operated in a pulsed mode. FIG. 11A illustrates changes in the
average fundamental and second harmonic generation (SHG) versus
current for pulses having a pulse length comparable to or longer
than a thermal time constant. Note that the average power is the
average over many pulses. Due to device heating, the fundamental
power output will decrease during an individual pulse. FIG. 11B
illustrates an example in which the pulses are significantly
shorter in temporal length than a thermal time constant. In this
case, the SHG output increases with current over a larger range of
currents and also results in a higher average power. The reason for
the improvement in SHG output with narrower pulses is that second
harmonic power output depends on the square of the fundamental
power. A consequence of the this square-law dependence on
fundamental power output is that selection of comparatively narrow
pulses can result in the situation in which the time-averaged
second harmonic generation is dramatically improved.
[0105] In a preferred embodiment, the pulsed laser source is
comprised of one or more surface emitting semiconductor lasers with
an extended cavity as described in patents by Mooradian ("High
power laser devices," U.S. Pat. No. 6,243,407; "Efficiency high
power laser device," U.S. Pat. No. 6,404,797; "High power laser,"
U.S. Pat. No. 6,614,827; "Coupled cavity high power semiconductor
laser," U.S. Pat. No. 6,778,582) which are incorporated here by
reference, with a non-linear optical material in the extended
cavity for converting the fundamental, infrared light from the
laser into visible light for display. In another embodiment the
source is formed from arrays of such lasers as described elsewhere
in this patent application.
[0106] Semiconductor lasers can also be driven in a color
sequential fashion at sub-frame speeds, such that a different color
can be on between any two, sequential sub-frames. FIGS. 12A and 12B
illustrate examples in which the light for a single pixel at the
projection screen for a system with a fast, binary modulator as a
function of time during the formation of one frame of an image. In
these examples the control electronics for driving the lasers is
synchronized with the control electronics for the spatial
modulator. Consequently an integer number of laser pulses is
generated during each time interval corresponding to a sub-frame of
the spatial light modulator. Additionally, the control electronics
determines the period of the color sequence, i.e., how many
back-to-back sub-frames are illuminated by laser pulses of a
particular color before the sequence moves on to the next color.
FIG. 12A illustrates an example in which the sequence has a small
period of sub-frames (e.g., the color moves in sequence with each
new sub-frame "sub-frame color sequential operation"). FIG. 12B
illustrates an example in which the sequence has a large period
(e.g., laser pulses of a particular color are generated for each
sub-frame over a larger group of sub-frames, such as over many or
all of the color sub-frames of a frame). Note that the control
electronics may be designed to have different modes of operations
such that one light engine can be adapted to change the period of
the color sequence depending upon the application.
[0107] In FIG. 12A, the laser pulses are presented in sequential
format without grouping at a rate faster than that capable with a
conventional white light source and color wheel. In particular, the
sequential format may be chosen to correspond to the sub-frame
level. This allows for the entire system to run in a color
sequential fashion at the sub-frame level, instead of at the frame
level. This means that any interruption in the light-path during a
frame results in a color bias level of merely one sub-frame.
Alternatively, the sequential format may be chosen such that the
system runs in color sequential fashion for a comparatively small
integer number of sub-frames, such that interruption of the
light-path during frame results in a color bias of merely a few
sub-frames. Since there are typically 2.sup.6 to 2.sup.10 possible
sub-frames (depending upon the number of switching periods), the
reduction in bias should be sufficient to render the problem
practically undetectable. In this manner, interruption of the light
path is perceived only as a reduction in intensity, and so a system
with a single spatial modulator 1010 can be perceived as performing
as well as the more complicated, multiple-modulator system.
[0108] Note that it is impractical to use a color wheel to obtain
sequential formats at a fine level of granularity. While it might
be possible in theory to use a color wheel and a white light source
to obtain color-sequential operation at the sub-frame level, it
would necessitate either extremely high rotation speeds (as high as
1 MHz), or a complicated color wheel with many filters disposed
about its edge. This latter approach is also extremely inefficient,
due to the blanking issue previously mentioned. Having many filters
means a similar number of blanking periods, which in turn reduces
the effective transmission of the color wheel, and increases the
required power from the light source.
[0109] In FIG. 12B, the light reaching the screen for a particular
pixel is composed of series of pulses of red, green and blue light,
where the colors are grouped and the groups of pulses are presented
sequentially. In some cases, pulses are skipped, demonstrating how
the total light reaching the screen of a particular color for a
single image can be modulated, i.e. the more pulses (ON states)
that are present, the brighter that particular color is for that
pixel in that frame.
[0110] The method of operation illustrated in FIG. 12B improves the
average power level. As previously described, the light sources may
comprise frequency doubled semiconductor lasers. One of the
fundamental aspects of non-linear frequency conversion of infrared
lasers is the supra-linear dependence of converted power on the
input power. A consequence of this is that pulsing the fundamental
semiconductor laser pump source can lead to enhanced output on an
average basis. This can be true even if the average power delivered
by the fundamental source is reduced from its CW average level. If
the dependence between input power and frequency converted power
goes as the input fundamental power raised to some exponent n,
where n>1, then the enhancement factor for pulsing over CW is
given by the ratio of the peak pulsed fundamental power to the CW
fundamental power raised to the nth power, then multiplied by the
fraction of the time that fundamental is on during one pulse cycle.
Thus, if the laser source operates in a QCW mode with each color
being capable of being separated in time via proper timing the
operating mode illustrated in FIG. 12B allows for the full average
power of the laser source to be used, while operation as in FIG.
12A allows only a fraction of the average power to be used. Bus the
method of FIG. 12B thus makes more efficient use of the pulsed
laser light.
[0111] Another advantage a light engine that supports both the
method of FIGS. 12A and 12B is that the laser can be used with
different types of spatial modulators, including spatial modulators
that are comparatively slow. Additionally, it is desirable that the
light engine be compatible with different systems such that the
light engine may be deployed in the highest volume possible.
[0112] Due to their reduced cost and complexity, color sequential
systems are very prevalent. This means that the requirements of
these systems have been designed into other, ancillary systems,
such as the electronic systems. In designing a new light source to
replace the white light source used in color sequential systems, it
is important to recognize these requirements and design the light
source accordingly, so that the new source may be rapidly adopted.
In particular it is desirable to have a light source and controller
such that the light engine may used as a generic replacement for
conventional UHP lamps in a wide variety of display
applications.
[0113] While light engines with digital spatial light modulators
have superior display characteristics due to their high speed, they
do not represent the entire market for light engines however.
Consequently it is desirable that the laser light source is
designed to be used in light engines having spatial modulators that
are slower but have a wide range of modulation, such as liquid
crystal on silicon (LCOS) modulators that can be adjusted from near
zero reflectivity (full-off) to a much higher reflectivity
(full-on). Such spatial modulators typically maintain a constant
level of modulation during a frame or sub-frame. In some designs,
one spatial modulator per color is used and the modulators change
level on a per-frame basis. In other designs, a single spatial
modulator is used and the color striking the modulator is changed
in sequence (color sequential operation). When the light source
emits white light, a color wheel is used to provide sequential
filtering of the light source. Such color wheels have, by nature,
borders between different filters, and during the time in which one
of these borders sweeps across the light source, the spatial
modulator must be turned off, so as not to provide distorted colors
to the viewer. These times are referred to as blanking periods.
[0114] FIG. 13 illustrates a method of driving a single color laser
source for color sequential operation compatible with light engine
designs having comparatively slow spatial modulators. In this
example, each color laser is on for about one third of the frame
times exclusive of a blanking period. Thus in this example each
laser of a multi-color laser source would be driven in a sequence
compatible with a conventional (comparatively slow) color sequence
that includes blanking periods. By making the operation of the
lasers compatible with existing systems the light engine becomes
useful as a replacement for conventional white light sources used
with comparatively slow non-DLP spatial modulators, such as LCOS
modulators. It will be understood that a single laser controller
may be programmed to have a dual-mode operation in which both DLP
(fast digital spatial modulator) and LCOS (slow spatial modulator)
are supported. For the case of comparatively slow spatial
modulator, visual artifacts are more noticeably (due to the longer
period of time that one laser color is active) and hence the method
of operation should minimize the output of lasers not actively
driven to imperceptible levels while also having the laser
"snap-on" to the desired power level almost immediately when the
lasers are active. These characteristics make it feasible to use a
multi-color laser source in a system with slow spatial modulators
and without a color wheel.
[0115] Referring again to FIG. 13, in a color-sequential display
system that utilizes a laser light source to illuminate
comparatively slow spatial modulators, it is also desirable to
reduce the output of all but one of the laser colors to
imperceptible levels during the periods where other colors are
being displayed, eliminating the need for a color wheel. The base
level that is imperceptible may be calculated for a particular
system by determining a maximum level of color contamination, e.g.,
how much blue and green light is acceptable when red is the
intended color, and so on. It may not be desirable to reduce the
base level drive of each laser source to zero, due to issues such
as warm-up time. Instead, the drive scheme can utilize the
threshold feature of a laser source to keep the drive level at a
thermally significant level while keeping the visible light output
at a minimal level, i.e., a drive level corresponding to an
optically quiescent off-state state where the laser produces
visible light below a pre-selected threshold level but is still
driven at a thermally significant level. The laser source may be
designed to take advantage of this scheme by incorporating one or
more design elements that boost the threshold of a laser, such as
laser cavities stabilized by thermal lensing. Additionally, if the
visible light is generated via non-linear processes, e.g. second
harmonic generation (SHG), then the non-linear relationships
between input drive and output light may be used to enhance the
aforementioned effect.
[0116] As illustrated in FIG. 13, an overdrive period may be
included during a blanking period. The blanking period may be used
to drive a laser source in such a way as to rapidly bring it into
its operating condition, without regard to the light output of the
laser source. In particular, if the laser has been running at a
reduced drive level during its off-state ("under-drive"), as
described in the preceding section, then the laser might be
over-driven during the blanking period to bring it back to normal
operating temperatures, and then brought back to normal drive
levels before the blanking period ends.
[0117] In many applications, it is desirable to provide a power
stabilization scheme for the laser. A conventional way to
accomplish this is to build a feedback loop based on power-current
characteristics of the laser. Note that the drive modulation may
also be adapted for use with laser arrays. For example, in a laser
array the laser array may be electrically configured into different
segments that are independently operable. Thus, a subset of the
lasers of a particular color may be pulsed at any one time. The
pulse drive may be applied in different spatial segments of the
laser array. The segmented pulsing may be selected to reduce
coherent locking. Additionally, the segmented pulsing may be
selected to stabilize the output power. However, in a preferred
embodiment with pulsed operation of a laser array, additional
flexibility in the power stabilization scheme is available. In the
case of a conventional power-current loop, an additional advantage
of an array is that the averaging provided by the array will lead
to more stable and uniform power-current relationships than that of
a single emitter. Another way to stabilize the power is to operate
with a fixed pulse current, but vary the pulse width and/or the
repetition rate. Yet another way to stabilize or vary the power is
to selectively vary the power from a part or the array, even to the
point of turning off individual elements. This can be achieved via
a properly designed electrical scheme.
[0118] Such schemes can also be used to drive laser sources for
display systems in which the spatial modulator is a fast beam
scanner, such as a pair of mirrors mounted on galvanometers. These
systems require separate intensity modulation that must occur at
the pixel modulation rate (video frame rate multiplied by the
number of pixels per frame. Direct-electrically pumped laser
sources, such as semiconductor lasers, are capable of being
directly modulated at these rates. In order to maintain image
fidelity, it is important to be able to accurately produce a given
intensity level. As mentioned above, the use of a laser array will
make this requirement easier to meet, as the power-current
relationship is more uniform due to averaging. Additionally,
modulation can be achieved by varying the pulse width, pulse
repetition rate and the current to individual elements. In a
preferred embodiment, the infra-red source is a surface emitting
semiconductor laser or laser array with an extended cavity and the
non-linear material is located inside the extended cavity and is
configured for second harmonic conversion, and the laser or laser
array is excited via direct electrical injection.
III. Improved Frequency Stabilization of Vertical Extended-Cavity
Surface Emitting Lasers and Laser Arrays with Nonlinear Frequency
Conversion
[0119] A. Thin Film Interference Filter
[0120] Referring to FIG. 14, an extended cavity laser may include a
gain element 1405, a spectral filter 1410, and output coupler 1420.
Wavelength (or, equivalently, frequency) control is an essential
part of laser design. While the general neighborhood of a laser
wavelength is defined by the material gain properties of the laser
system, it typically takes more effort to tailor the wavelength
output of the laser for a particular application.
[0121] The wavelength control of semiconductor lasers is a common
problem for laser designers since the optical gain of semiconductor
materials typically spans tens of nanometers in the wavelength
space while the final application may require sub-nanometer
wavelength spectrum.
[0122] Intracavity frequency-doubled vertical extended-cavity
surface-emitting lasers (VECSELs) have been demonstrated to be
useful in a variety of applications. An intracavity
frequency-doubled VECSEL includes a surface emitting gain element
1405 and an output coupler 1420 spaced apart from the surface
emitting laser to form an extended cavity. The output coupler 1420
reflects light generated at the fundamental frequency back towards
the surface emitting gain element 1405. An intracavity frequency
doubling crystal 1415 generates light at a frequency equal to twice
the frequency of the fundamental beam.
[0123] The conversion efficiency of an intracavity
frequency-doubled VECSEL depends upon the longitudinal mode
characteristics of the VECSEL at the fundamental frequency.
Nonlinear frequency conversion processes, such as second harmonic
frequency doubling, are typically wavelength (frequency) dependent
with a typical bandwidth on the order of 1 nm or less for most
nonlinear crystals of practical length (e.g., several millimeters).
Consequently, in an intracavity frequency doubled VECSEL control of
the frequency and longitudinal mode spectra of the laser light
about the fundamental frequency is crucial to obtain a high output
power of frequency doubled light. Depending upon the application,
control of the frequency within a selected range improves
non-linear conversion efficiency. In the context of a light engine,
the frequency range can be further selected to be consistent with a
range of spectral pulse broadening to reduce speckle.
[0124] Additionally, the conversion efficiency of an intracavity
frequency-doubled VECSEL depends critically upon the optical losses
within the extended cavity. It is generally known that optical
losses reduce the efficiency of intracavity frequency doubling
processes. See, e.g., R. Smith, "Theory of intracavity optical
second-harmonic generation," IEEE Journal of Quantum Electronics,
vol. 6, p. 215, (1970), the contents of which are hereby
incorporated by reference. Although several effects occur,
frequency doubling processes are nonlinear processes in which the
efficiency of the conversion process depends nonlinearly on the
circulating power at the fundamental wavelength. In order to build
up a high circulating power at the fundamental wavelength in a
VECSEL comparatively low losses in the extended cavity are
required.
[0125] Calculations by the inventor of the present patent
application indicate that frequency-doubled VECSELs require
especially low losses in order to achieve high conversion
frequencies. In particular, calculations by the inventor of the
present patent application indicate that a 1% single-pass loss at
the fundamental wavelength (frequency) of the surface emitting gain
element 1405 due to an intracavity spectral filter 1410 can easily
result in over 10% loss in the intracavity circulating power at the
fundamental wavelength, which results in a decrease in the
second-harmonic power (for intracavity frequency doubling) by
15-to-20% or more.
[0126] The design of a spectral filter for a VECSEL with
intracavity frequency doubling involves a consideration of several
tradeoffs. Spectral filters capable of controlling the frequency of
the intracavity frequency doubled VECSEL also tend to introduce an
associated optical loss. Inserting additional optical elements into
an intracavity frequency-doubled VECSEL to control the frequency
results in a tradeoff between the increased conversion frequency
provided by controlling the fundamental frequency and the decrease
in power associated with increased optical loss. Only frequency
selective elements with a small optical loss result in a net
improvement in conversion frequency of an intracavity
frequency-doubled VECSEL. Typically the constraint on optical loss
again has limited the choice of frequency selective elements in the
wavelength range of typical semiconductor lasers to coated etalons
and birefringent filters, which can have losses of about 1% or
less. By way of contrast, notch filters commercially available from
Edmunds Optics have their maximum transmission specified as about
90%, which is unsuitable for intracavity laser applications.
Background information on etalons and birefringent filters is
described in C. C. Davis, "Lasers and Electro-Optics: Fundamentals
and Engineering," Cambridge University Press, 2002, p. 73) and P.
J. Valle and F. Moreno, "Theoretical study of birefringent filters
as intracavity wavelength selectors," Applied Optics v.31, p. 528
(1992), the contents of each of which are hereby incorporated by
reference. A typical arrangement consists of a spectral filter
(Fabry-Perot etalon or birefringent) positioned in the laser cavity
at an angle to the laser optical axis defining the beam direction.
This tilt angle is typically used to suppress unwanted feedback
effects and/or to angularly tune the spectral filter to the desired
spectral performance.
[0127] However, while both Fabry-Perot etalons and birefringent
filters can be selected that have losses less than about 1%, these
spectral filters do not provide the desired degree of control over
frequency and also have manufacturing drawbacks that may limit
their manufacturability. As one example, Fabry-Perot etalons have
an optical response in which there are peaks in transmissivity that
have a periodicity determined by the optical thickness of the
etalon and the wavelength of light. In the context of the gain
spectrum of laser diodes this means that a single etalon may not
provide sufficient longitudinal mode discrimination over an
extended range of operating conditions. In particular, a number of
transmission peaks of the etalon may lie within the gain spectra of
the laser diode such that the longitudinal mode is not locked over
all operating conditions.
[0128] As an example of some of the problems associated with using
etalons to lock the frequency of an extended cavity surface
emitting semiconductor laser, consider a practical case of
designing an etalon filter with a bandwidth (defined as the full
width at half-maximum, or FWHM) of 0.4 nm near 1064 nm central
wavelength. Such a bandwidth can be approximately achieved, e.g.,
with a fused silica etalon of 300 micron thickness with both faces
coated for about 35% reflectivity at a 1064 nm wavelength. A
corresponding transmission curve is illustrated in FIG. 15. While
the spectral transmission peak centered at 1064 nm satisfies a
desired bandwidth goal of 0.4 nm, the neighboring transmission
peaks located at approximately 1.3 nm (etalon free spectral range,
or FSR) from the central peak may provide undesirable spectral
channels for laser emission. Since a semiconductor laser may have a
gain spectra that extends over tens of nanometers it can be
understood from FIG. 15 that many different etalon peaks would be
within the gain spectra of the laser. Consequently, the frequency
discrimination provided by the etalon may be insufficient.
[0129] One possible solution to this problem would be to use
thinner etalons to increase the spectral range such that fewer
etalon peaks would be within the gain spectra of the laser. A
thinner etalon has a wider separation between resonant peaks but
also requires higher reflectivity to achieve a narrow bandwidth.
However, a thin etalon with high reflectivity mirrors can lead to
higher walk-off diffraction losses for Gaussian beams transmitting
through a tilted etalon. Additionally, very thin etalons (under 100
micron) become more difficult to manufacture, coat, and handle.
[0130] Another possible solution can be achieved by utilizing two
etalons such that the combined optical response has a narrow
bandwidth and a larger separation between peaks in transmissivity
than a single etalon. An intracavity laser with two etalons is
described in the U.S. patent application Ser. No. 10/745,342
"Compact Extended Cavity Laser" by C. A. Amsden, M. K. Liebman, A.
V. Shchegrov, and J. P. Watson. However, a two-etalon laser design
adds complexity and cost to the laser.
[0131] Birefringent filters have somewhat similar problems to those
of etalons. The most common material used for intracavity
birefringent filters is crystalline quartz. Producing a spectral
filter with a FWHM of 0.4 nm would require a thick piece of quartz
(on the order of 1 cm or thicker depending on the crystal cut and
angular orientation in the cavity) and this might make it
impractical to use in a compact and low-cost cavity. One solution,
again, is to use several birefringent filter plates but it also
adds complexity and cost undesirable for a compact and low-cost
laser.
[0132] Yet another consideration is that both etalons and
birefringent filters rely on interferometric effects with a large
variance in the resonant frequencies of manufactured filters. For
example, to predict the exact resonance frequency of an etalon one
would have to manufacture the thickness of the etalon with an
accuracy within a fraction of a wavelength. The manufacturing
variance in filter response of etalons can be tuned to match the
spectral filter peak wavelength with the peak wavelength of the
nonlinear material, using mechanical, thermal, or electro-optical
tuning. However, these options frequently require significant
excursions (e.g., angular or thermal) from desired nominal values
and may be undesirable in a laser system that has to be designed
for compactness and low cost.
[0133] Manufacturing considerations also tend to limit the design
of etalons and birefringent filters in a variety of ways. Etalons
have a spectral range (separation between resonant peaks) that
depends inversely on the thickness of the etalon. For example, as
previously described, a thin etalon has a large spectral range).
However, a thin etalon also has a larger percentage variation in
the position of a spectral filter peak with thickness than a
thicker etalon. Note also that etalons will have manufacturing
tolerances that vary from etalon-to-etalon and also manufacturing
tolerances with respect to the thickness along one etalon. As a
consequence, it is difficult to manufacture large area etalons
capable of supporting arrays of surface emitting lasers covering a
large area.
[0134] In light of the above-described problems, the apparatus and
system of the present invention was developed. Embodiments of the
present invention describe methods for stabilizing the frequency of
surface-emitting lasers or laser arrays using a frequency selective
filter that replaces conventional frequency selective filters, such
as etalons and birefringent filters. This provides potential cost
reductions, improves manufacturability, allows for an accurate
design of the output laser wavelength, and does not require costly
mechanical or thermal tuning of the spectral filter to achieve
desired spectral properties. Embodiments of the present invention
are also especially suited for frequency stabilizing lasers or
laser arrays designed for intracavity nonlinear frequency
conversion, such as second harmonic generation using engineered
periodically poled nonlinear materials. This is especially
important because such nonlinear materials can be engineered for
nonlinear conversion at exactly the target wavelength and having a
simple and low-cost approach to frequency stabilize the laser
without extra mechanical or thermal tuning, resulting in
significant cost advantages over other alternative approaches for
frequency stabilization, such as etalons and birfingrent
filters.
[0135] FIG. 16 illustrates a VECSEL with intracavity frequency
doubling in accordance with one embodiment of the present invention
in which conventional spectral filters are replaced with a
thin-film interference filter 1610. The VECSEL has a surface
emitting semiconductor gain element 1605 disposed on a chip for
generating light at a fundamental frequency. An exemplary gain
element is based on the design described in patents by A. Mooradian
("High power laser devices," U.S. Pat. No. 6,243,407; "Efficiency
high power laser device," U.S. Pat. No. 6,404,797; "High power
laser," U.S. Pat. No. 6,614,827; "Coupled cavity high power
semiconductor laser," U.S. Pat. No. 6,778,582) which are
incorporated here by reference. Numerous variations of the basic
gain element 1605 are within the scope of the present invention,
such as variations in the Bragg mirror structure and the
incorporation of a lens or lenses to stabilize the spatial
eigenmode.
[0136] A nonlinear crystal 1615 may optionally be included. An
output coupler 1620 spaced apart from gain element 1605 defines an
extended cavity. Output coupler 1620 may include, for example, a
mirror to reflect light at a fundamental frequency back to gain
element 1605.
[0137] In the preferred embodiment, the gain element 1605 contains
two epitaxially grown quarter-wavelength Bragg mirror stacks 1601
and 1603, one of which is highly reflective at the design
wavelength, e.g. 976 nm, and another is that partially reflective
at this wavelength. The gain medium 1607, which typically comprises
quantum wells, is also epitaxially grown and is sandwiched between
the two mirror stacks in a region 1609 having an optical thickness
at the fundamental wavelength of corresponding to a whole number of
wavelengths. The arrangement of Bragg mirror stacks 1601 and 1603
spaced apart by a resonant spacer layer results in gain element
1605 having a resonant frequency and associated bandwidth. In
practical applications, however, the real and imaginary components
of the refractive indices vary in the gain region due to the
presence of optical gain and electron-hole pairs. As a consequence,
Bragg mirror stacks 1601 and 1603 do not, by themselves, provide
sufficient control of the laser wavelength over a range of
operating conditions.
[0138] In an intracavity frequency doubling configuration the
extended cavity laser is preferably designed to build up a large
density of photons at the fundamental frequency. For example, the
output coupler 1620 may be designed to have an extremely high
reflectivity about the fundamental frequency such that photons at
the fundamental frequency make many round-trips within the cavity.
This circulating light at the fundamental frequency is partially
converted to light at the frequency-doubled frequency in each pass
through the nonlinear crystal 1615. However, nonlinear conversion
processes are extremely sensitive to input power densities.
Consequently, in an intracavity frequency doubling configuration it
is desirable to design the extended cavity laser to have a high
power density of light at the fundamental frequency circulating
within the extended cavity. Light at the frequency-doubled
frequency is coupled out of the cavity, either directly through the
output coupler or via an additional coupling mechanism (not shown)
to selectively couple light at the second harmonic out of the
extended cavity.
[0139] In one embodiment thin film filter 1610 includes a Bragg
mirror configuration selected to provide additional frequency
control. In a preferred embodiment, the thin film interference
filter 1610 includes Bragg mirrors defining a resonant response. In
one embodiment, thin film interference filter 1610 has a resonant
response that is an optical analog to that of the gain element
1605. In particular, the Bragg mirrors and separation between Bragg
mirrors may be an optical analog to the Bragg mirror structure of
gain element 1605. In this embodiment, one essentially separates
part or all of the wavelength control function from the gain
element 1605 and moves it into the external cavity.
[0140] FIG. 17 illustrates in more detail an exemplary thin film
interference filter 1610 in accordance with one embodiment of the
present invention. The thin film layers 1701, 1703, and 1709 are
formed on a substrate 1720 using conventional deposition
techniques. While different substrates can be used including
semiconductor materials such as GaAs, it may be convenient to use
common optical substrates such as fused silica. The substrate and
the thin film coatings may be selected to have superior control and
stability of refractive indices compared with semiconductor
materials. For example, all of the optical layers and thin films
may be formed from materials which are not semiconductors, such an
insulators and/or metal oxides. An exemplary structures includes
two stacks 1701 and 1703 of quarter-wave higher- and lower-index
layer pairs and a resonant spacer 1709 (which can also be fused
silica) between mirror stacks. The mirror layer stacks 1701 and
1703 may have identical reflectivity and a non-resonant spacer
layer 1711 may be grown on the second mirror stack to ensure the
identical reflectivity of the two stacks. Finally, the outer layers
on both sides of the structure are designed to be anti-reflective
for the target wavelength (here, 976 nm), and, if required, for the
second-harmonic wavelength (488 nm). In contrast with conventional
solid etalons, which are fabricated from, e.g. fused silica or BK7
optical glass thinned to a prescribed thickness and coated on both
sides to the same reflectivity specification, the interference
filter design of the present invention allows for an accurate
design of the target maximum transmission wavelength.
[0141] FIG. 18 illustrates in more detail an exemplary sequence of
thin film coatings to implement an interference filter. In this
example, silicon dioxide and tantalum oxide coatings are deposited
on a fused silica substrate using, for example, ion beam
sputtering. The coatings are designed for a resonant transmission
peak for the fundamental wavelength of 976 nanometers with a FWHM
of 0.4 nm. Note that the thin film interference filter will also
transmit light at the second harmonic of 488 nm (since the
quarter-wave stacks will appear as half-wavelength stacks for the
second harmonic frequency having half the wavelength of the
fundamental). Dual band antireflection coatings are deposited on
both sides of the interference filter.
[0142] FIG. 19 illustrates theoretical calculations of the optical
transmission response of the filter structure of FIG. 18 modeled
with a thin-film design software TFCalc. The theoretical
transmission peak at the fundamental target wavelength approaches
100% with a FWHM of 0.4 nm.
[0143] FIG. 20 is a plot illustrating empirical studies of a
fabricated filter having a layer sequence similar to that of FIG.
18. The interference filter was fabricated using the ion-beam
sputtering (IBS) coating technique. The transmission measurement
for this filter was then performed using a tunable laser. The FWHM
is 0.4 nm and the transmissivity at the target wavelength is
greater than 99%. Note also that there is only one resonant peak
within an extended wavelength range greater than several
nanometers. In particular there will be only one resonant peak
within a wavelength range of conventional semiconductor gain
materials, which can have gain over a wavelength range of ten
nanometers. Thus, the resonance response will facilitate locking
the frequency to a desired wavelength range, unlike conventional
etalons that have a limited spectral range.
[0144] An alternate design for an interference filter is to use
slightly different recipes for the "left" and "right" Bragg mirror
stacks with the structure optimized to equalize their
reflectivities and create an anti-reflection effect for the second
harmonic (488 nm) for the "right" mirror stack. This design can
eliminate the non-resonant spacer layer and the anti-reflection
coating stack on the right making a simpler coating.
[0145] Stabilizing a laser or a laser array with a thin-film
interference filter of the type described above was not recognized
as a practical solution until the design described in this
invention was proposed by the inventor and a test was performed.
There are several reasons for this. First, while narrowband
spectral filters with the bandwidths illustrated (also known as
notch filters) have been designed and commercially fabricated for a
range of applications they are generally not as narrowband as the
filters shown in FIGS. 19-20 and have significantly higher
transmission losses. For example, conventional notch interference
filters from Edmonds have a transmission of only about 90%. In
contrast, as seen in both FIG. 19 (model), and, somewhat
surprisingly, in FIG. 20 (experiment), the transmission of
interference filters formed in accordance with the teachings of the
present invention exceeds 99% with a FWHM less than one nanometer,
which is what an intracavity laser requires for many
applications.
[0146] The thin-film interference filter 1610 is positioned in the
external cavity, preferably at an angle to avoid feedback into the
gain region. Thus, such a filter plays a role similar to both a
traditional coated solid etalon and to the Fabry-Perot etalon
combined with the laser gain chip, combining the advantages of both
cases. A nonlinear crystal 1615, properly selected for frequency
doubling (or, in general, nonlinear frequency conversion) is also
positioned in the extended cavity that is completed by the output
coupler 1620. The output coupler coating is preferentially selected
to be highly reflective at the fundamental wavelength (the design
wavelength of the interference filter) and highly transmissive at
the frequency converted wavelength. The cavity length, crystal
length, and the curvature of the output coupler 1620 are properly
optimized for the design goals as known in the art of laser design.
In the preferred embodiment, the nonlinear crystal is a
periodically poled nonlinear crystal such as periodically poled
KTP, LiNbO3, or LiTaO3. The advantage of using such crystals is
that they posses high nonlinearity and can be engineered for
nonlinear frequency conversion of the wavelength of choice.
[0147] The role played by the thin-film interference filter 1610 is
to narrow the laser emission to a desired bandwidth. In some cases,
such as required in instrumentation applications, the bandwidth of
the filter has to be narrow enough to ensure the
single-longitudinal mode operation of the laser. This will also
ensure a stable, low-noise laser operation and, when properly
tuned, will constrain the emission at the peak of the nonlinear
conversion bandwidth. In other cases, the bandwidth of the
thin-film interference filter may be designed to be broader so that
the laser can operate in more than one longitudinal mode to achieve
desirable range of frequencies. However, it is still desirable to
constrain these modes within the nonlinear conversion bandwidth of
the nonlinear crystal 1615.
[0148] The nonlinear crystal 1615 may be used to control the
polarization of the surface-emitting laser as described in the U.S.
patent application Ser. No. 10/734,553, "Polarization control of an
extended cavity laser" by A. V. Shchegrov, A. Umbrasas, J. P.
Watson, and D. Lee, which is incorporated here by reference.
[0149] Furthermore, to improve compactness and decrease the cost of
the laser, it may be desirable to deposit the thin-film
interference filter 1610 on one of the crystal faces of the
nonlinear crystal 1615 and combine two elements in one unit. Such
an arrangement is another embodiment of this invention.
[0150] One important aspect of the design of interference filter
1610 is that the design may be an optical analog to the wavelength
control elements of the surface-emitting laser chip 1605. For
example, the optical thicknesses of layers in the surface-emitting
laser chip can be used as starting point in designing a thin film
interference filter design having a well-defined resonance about
the fundamental frequency. As is well known in optics, the optical
thickness of a layer depends upon the optical wavelength, which in
turn is a function of the frequency of the light and the refractive
index of the medium that the light is passing through A surface
emitting gain element 1605 has Bragg reflectors 1601 and 1603 that
are typically designed to product a resonant response at a
fundamental frequency. The resonant response will also have an
associated bandwidth. The design of the surface emitting gain
element (e.g., number of layers in Bragg stacks) is partially
constrained by other design considerations, such as achieving a
high optical power in the extended cavity and reducing material
losses within the gain element. As previously mentioned, the
surface-emitting chip itself provides some degree of frequency
control but tailoring it for a narrow enough bandwidth may lead to
the loss of power that is available in the external cavity for
frequency conversion. In one embodiment the design of the
interference filter is based on the design of the surface-emitting
laser chip (converted to a thin film implementation), tailored to
produce a desired frequency selection, and implemented as an
intracavity spectral filter with high peak transmission. Obviously,
there are several significant differences between the interference
filter and the laser chip. One difference is that the absence of
the gain (active) layer in the interference filter. Another
difference is that the laser chip structure is usually based on
semiconductor materials such as GaAs, AlAs, composites like GaAlAs,
and others. This is done to create a high-quality structure that
can incorporate gain layers (quantum wells) and perform reliably
under stress, i.e. during the laser operation. For the interference
filter, the stresses are much less severe since this is not an
active element. Instead, the primary design goals are low losses
and certain spectral bandwidth. This is why the thin film
interference filter is preferably manufactured with dielectric
coating layers, preferably deposited with ion-beam sputtering (IBS)
method. This is done to ensure the lowest loss possible.
[0151] FIG. 21 illustrates an embodiment in which the thin film
interference filter 1610 is used to provide frequency control of a
surface-emitting, extended-cavity laser array 2105 of gain elements
formed on a common chip or substrate. A lens array may be
integrated with the surface emitting laser 2105 to control the
spatial mode for each emitter. Alternately, thermal lensing may be
used to control the cavity spatial mode for each emitter. As
illustrated in FIG. 21, the components illustrated in FIG. 21 are
all preferably segments with planar surfaces, which facilitates a
low-cost assembly with few critical alignments, as described below
in more detail. The partially reflective mirror integrated with the
laser chip may or may not be present. When this mirror is present,
its primary function is not the frequency selectivity for the laser
by establishing the Fabry-Perot etalon between itself and the
highly reflective mirror but separating the gain medium from losses
in the external cavity, e.g. losses in the substrate on which both
mirrors and the gain layers are grown. A lens or lenses may be
integrated with the chip or positioned separately in the extended
cavity. When integrated with the chip, the lens may be a thermal
lens generated by the heat created in the gain area or a static
lens, e.g. etched on the chip surface. Such a lens may help to
stabilize the spatial eigenmode of the laser and is especially
desirable when one wants to design the external cavity with a flat
output coupler for simplicity and/or cost reasons.
[0152] For an array embodiment, the output coupler 1620 is
preferably flat and it is common for all the emitters in the laser
array. However, it is also possible to use an array of
micro-mirrors forming curved output couplers for each individual
emitter. The same nonlinear crystal 1615 may be used to convert the
fundamental wavelength into another wavelength, such as second
harmonic (half of the fundamental wavelength).
[0153] In a low-cost array assembly, the interference filter may be
used as a separate element or attached to another components, such
as the nonlinear crystal 1615, to reduce the number of components
that must be optically aligned for the assembled laser.
Alternately, the interference filter also be combined with one of
the surfaces of the output mirror 1620, but in this case it is
preferable to wedge the output mirror element and use the inner
(intracavity) surface for the filter coating and the outer surface
for the high-reflection mirror coating.
[0154] While the interference filter may be used as the only
spectral filter, more generally it will be understood that the
interference filter may also be combined with other spectral
filtering elements to control the wavelength of the laser. As one
example, the interference filter may be attached, bonded, or
deposited on a wedged surface of an end mirror of the extended
cavity.
B. Volume Bragg Grating for Frequency Control
[0155] In accordance with one embodiment of the present invention,
a volume Bragg grating is utilized as a spectral filter for a
VECSEL with intracavity frequency doubling. As previously
described, VECSELs with intracavity frequency doubling have several
requirements for high conversion efficiency. First, the wavelength
must be locked within an optimum bandwidth of the nonlinear
crystal, typically less than about one nanometer such that the
resonance response has a FWHM of less than one nanometer. Second,
the spectral filter preferably has no secondary peaks within the
bandwidth of the semiconductor gain medium, i.e., there is a single
resonant peak over a wavelength range of several nanometers, such
as a ten nanometer range. Third, the total optical loss introduced
by a spectral filter at the fundamental wavelength is preferably
less that about 1%. In one embodiment of the present invention,
VBGs are used for frequency control of surface-emitting laser array
with intracavity frequency doubling.
[0156] Volume Bragg gratings (VBGs) are wavelength-selective
reflective elements that are made of special glass with a periodic
refractive index variation written in it. Such an index variation
can be designed to produce a spectrally narrow high-reflectivity
element that can help to control the spectrum of the laser in a
window selected by design. While fiber Bragg gratings have been
known for several years in telecom laser design applications at
telecom wavelengths (e.g., 1.55 microns), their volume counterparts
(VBGs) have been commercially available only fairly recently.
Manufacturing details and characteristics of such grating elements
are described in U.S. Pat. No. 6,586,141, "Process for production
of high efficiency volume diffractive elements in photo-thermal
refractive glass" by O. M. Efimov, L. B. Glebov, V. L. Smirnov, and
L. Glebova, and U.S. Pat. No. 6,673,497, "High efficiency volume
diffractive elements in photo-thermal refractive glass" by O. M.
Efimov, L. B. Glebov, and V. L. Smirnov. Similar holographic
elements have also recently been fabricated and used in media
storage technologies (e.g., by InPhase Technologies) and while we
will use the term volume Bragg grating (VBG) in the following
discussion, it is assumed that using such holographic grating
elements is also within the scope of this invention. Recently, it
was described how to use VBGs for frequency stabilization of
edge-emitting lasers and laser arrays (G. Venus, V. Smirnov, L.
Glebov, "Spectral Stabilization of Laser Diodes by External Bragg
Resonator", Proceedings of Solid State and Diode Laser Technology
Review, Albuquerque, N. Mex., June 2004, B. L. Volodin, V. S. Ban,
"Use of volume Bragg gratings for the conditioning of laser
emission characteristics," U.S. patent application Ser. No.
10/884,524).
[0157] Volume Bragg grating elements are now currently commercially
available from Optigrate (Orlando, Fla.), PD-LD (Pennington, N.J.),
and Ondax (Monrovia, Calif.). Empirical studies were performed to
determine whether commercially available VBGs were suitable for
frequency stabilizing vertical cavity surface emitting laser with
nonlinear frequency conversion.
[0158] Empirical investigations by the inventors of the present
patent application have demonstrated that VBGs may be designed to
have a combination of FWHM and low losses useful for VECSELs with
intracavity frequency conversion to generate green and blue light
from an infrared fundamental pump wavelength.
[0159] FIG. 22 is a plot of the reflectivity of a commercially
available VBG with wavelength. Reflectivity measurements were
performed with a tunable laser on a VBG designed at the target
wavelength of 976 nm. The empirical results demonstrate that a FWHM
of less than 1 nm, i.e., 0.6 nm. The reflectivity is in excess of
99%. Although the empirical reflectivity has slight irregularities
outside of the FWHM, the reflectivity drops off rapidly at the FWHM
with no satellite peaks above the FWHM. Thus the empirical results
demonstrate that there is only one dominant peak within a
wavelength range of conventional semiconductor gain materials,
which can have gain over a wavelength range of ten nanometers.
Thus, the empirical results demonstrate that a VBG can be designed
that has a combination of attributes suitable for a VECSEL with
intracavity frequency doubling.
[0160] FIG. 23 illustrates an embodiment of a VECSEL with
intracavity frequency conversion (e.g., frequency doubling)
utilizing a VBG 2310 to frequency-stabilize a surface-emitting
laser. VBG 2310 also serves as a reflector element of an output
coupler, thereby defining an extended cavity. The semiconductor
laser gain element 1605 and nonlinear crystal 1615 are described in
the embodiment illustrated above in FIG. 16.
[0161] A polarization control element is preferably included. In
one embodiment a polarizing beamsplitter 2320 provides polarization
control function and is also used to re-direct the back-propagating
frequency-converted beam (illustrated in phantom by line 2360)
outside of the cavity. To perform these functions, the two surfaces
of such a polarizing beamsplitter 2320 are coated to the provide
high transmission for the desired polarization at the fundamental
laser wavelength and one of the surfaces is coated for high
reflection at the frequency converted wavelength. In an alternate
embodiment polarization is controlled by the birefringence of
nonlinear crystal as described in U.S. patent application Ser. No.
10/734,553, "Polarization control of an extended cavity laser" by
A. V. Shchegrov, A. Umbrasas, J. P. Watson, and D. Lee.
[0162] In some application, it may be desirable to design the VBG
2310 for highest reflectivity at the fundamental design wavelength,
such as 920 nm, 976 nm, 1064 nm, or 1260 nm. Then both faces of the
VBG can be coated with anti-reflective coating at both the
fundamental wavelength and frequency-converted wavelength, e.g. 460
nm, 488 nm, 532 nm, or 630 nm. In other cases, it may make sense to
design a VBG primarily to achieve the spectral bandwidth target and
sacrifice some of the maximum reflectivity. To obtain a
high-efficiency intracavity nonlinear conversion, a VBG may be
coated by a highly reflective coating 2340 at the fundamental
wavelength to "close" the cavity and maximize intracavity
circulating power. Coating 2340 may also be designed to be
anti-reflective at the frequency converted wavelength (e.g. at the
second harmonic for frequency doubling) to couple
frequency-converted light out from volume Bragg grating 2310. This
permits a configuration in which for each pass of a fundamental
beam through nonlinear crystal 1615 in a forward direction toward
the VBG results in frequency converted light that may leave the
cavity through the VBG whereas light at the fundamental which is
reflected back into the extended cavity from the VBG travels in a
backwards direction through nonlinear crystal 1615 with the
resultant frequency-converted light coupled out of the cavity by
polarizing beamsplitter 2320.
[0163] The spectral properties of the VBG may be tailored for
particular applications. In some cases it is desirable to design
the bandwidth of VBG to be narrow enough to provide low-noise
single-longitudinal mode operation for a high-performance,
instrumentation-quality laser. In other cases, it may be desirable
to allow several longitudinal modes and constrain them, e.g. within
the spectral bandwidth matching the bandwidth of nonlinear
conversion process.
[0164] FIG. 24 illustrates an embodiment in which a VBG is used to
provide optical feedback and spectral filtering for a surface
emitting array. Using a VBG for frequency stabilization of
surface-emitting extended-cavity laser arrays is especially
advantageous. A single VBG element 2340 is used to frequency
stabilize the entire laser array of gain elements 2105 and a single
nonlinear crystal 1615 is used to frequency double all the surface
emitters in the arrays. A single beamsplitter 2320 is used to
define polarization and, if so desired, to extract the
frequency-converted back-propagating array of beams from the
cavity. The forward-propagating, frequency converted array of beams
may be extracted through the VBG. In another case, which is also
within the scope of this invention, a forward and a backward
propagating beams may be collected in the same direction if a
dichroic coating that is highly reflective for a frequency
converted light is placed on an appropriate surface of the
nonlinear crystal, VBG, or the semiconductor laser chip. It may be
desirable to spatially separate the two frequency converted beams
(forward and backward) for each emitter by placing a wedge on or
using a tilt of the appropriate optical surface.
[0165] In many applications desirable to operate all array elements
at substantially the same frequency to maximize nonlinear
conversion and minimize cost. However, in some cases it may be
advantageous to run different emitters at slightly different
frequencies. This could help in reducing cross-talk between
different elements and in reducing coherence of the entire laser
array source, which can help in reducing speckle from such a
source.
[0166] Conventional VBG applications are typically single
wavelength. However, in the present invention the VBG is preferably
modified to take into account that the system includes a
fundamental (pump) wavelength and a frequency converted wavelength
(e.g., the second harmonic wavelength). One aspect is appropriate
design of optical coatings on the VBG, as previously described.
Additionally, the preferred design of this invention for VBG is
dichroic, i.e. the VBG has to have narrow-spectral-brand
reflectivity at the fundamental laser wavelength and non-resonant
high transmission at the second harmonic. In one embodiment, the
VBG's dichroic attributes are optimized by theoretical and/or
empirical modeling. For example, several different VBG grating
attributes and compositions may be tested for a particular choice
of fundamental wavelength and second harmonic wavelength to
optimize the dichroic properties of the VBG for a particular
application. Empirical testing was performed and demonstrated that
a VBG can be optimized to have suitable dichroic properties for
frequency stabilization of surface emitting laser with nonlinear
frequency conversion.
[0167] In some embodiments, it may be desirable to monolithically
bond the VBG and nonlinear crystal. Normally, this will not affect
phase-matching or quasi-phase-matching thermal tuning of the
nonlinear crystal because VBGs typically possess very high
temperature stability of their reflection spectrum with spectral
tuning of about 0.01 nm/.degree. C.
[0168] Note that the laser designs described above are not
restricted to any particular mode of operation and may be
continuous wave or pulsed, e.g. pulsed by direct current pulsing of
the electrically pumped semiconductor surface emitters. In
addition, surface-emitting lasers described in this invention can
be mode locked using a saturable absorber element (see, e.g., K.
Jasim, Q. Zhang, A. V. Nurmikko, A. Mooradian, G. Carey, W. Ha and
E. Ippen, "Passively modelocked vertical extended cavity surface
emitting diode laser", Electronics Letters, V.39, p. 373 (2003)).
In this case, a thin-film interference filter or a VBG can be
designed to control the spectral width of the mode-locked output to
optimally match the spectral phase-matching bandwidth of the
nonlinear crystal to optimize the speckle reduction per emitter.
All the designs described above are scalable to array architecture
in both one and two dimensions but they can be equally applied to
single-emitter lasers. Finally, surface-emitting lasers and laser
arrays frequency stabilized by the methods described in this
invention may be either electrically or optically pumped.
[0169] In one of primary application embodiments of this invention,
the surface-emitting gain elements are designed around the RGB
wavelengths that are double of the wavelengths of interest to
projection display applications, i.e. .about.635 nm (red),
.about.532 nm (green), and .about.460 nm (blue).
[0170] Another embodiment of this invention is frequency
stabilization of frequency doubled lasers or laser arrays to make
them operate in an extended range of ambient temperatures without
costly active temperature control (e.g., temperature control by
thermo-electric coolers). In many applications, involving, e.g.
handheld devices such as PDAs, laser pointers, etc., it is
desirable to design them for operation in a wide range of ambient
temperatures, e.g. -10.degree. C. to 50.degree. C. Conventional
solutions would require active temperature control of the laser and
critical components, e.g. by thermoelectric coolers. The present
invention can take advantage of the fact, mentioned before, that
volume Bragg gratings are typically very temperature stable
elements, tuning at 0.01 nm/.degree. C. or less, and they can be
further designed for temperature insensitivity. This allows locking
the wavelength of the laser at the desired value and the nonlinear
crystal can be tuned to the locked laser wavelength by an
inexpensive heater element designed in the package. This solution
may be used on different types of lasers or laser arrays including
surface-emitting and edge-emitting semiconductor lasers and
solid-state lasers, which may be pumped by different means.
[0171] There are several types of applications where
frequency-stabilized lasers or laser arrays can be especially
useful. Among them are applications as visible light laser or laser
array sources for projection displays, instrumentation applications
of single-longitudinal mode lasers such as flow cytometry or
confocal microscopy, printing applications, lighting applications,
and others.
IV. Highly Manufacturable, Compact Frequency Doubled Vertical
Extended-Cavity Surface Emitting Laser Arrays
[0172] Embodiments of the present invention describe a technical
approach of constructing visible laser arrays in a scalable and
low-cost platform. The scalability applies to both the output power
of the array and to the price that can be scaled down in volume,
wafer-scale manufacturing. Furthermore, this technical approach
allows obtaining all desirable visible colors in similar fashion
again resulting in cost reduction when manufacturing such RGB laser
arrays.
[0173] One embodiment of the present invention describes low-cost,
highly manufacturable designs for frequency-doubled arrays of
frequency-doubled vertical extended-cavity surface-emitting lasers
(VECSELs). In a VECSEL system, the laser cavity is defined by the
surface-emitting laser chip upon which an array of surface emitting
gain elements are formed and the end mirror. The laser chip has at
least one epitaxially grown mirror and a gain region. In a
preferred embodiment, the gain region comprises one or several
quantum wells.
[0174] FIG. 25 is a block diagram illustrating major components of
a VECSEL array. A surface emitting laser gain chip has an array of
gain elements 2510. A polarization control element 2520 is provided
to control the polarization. A nonlinear crystal 2530 is provided
for frequency doubling. A wavelength control element 2540 is
provided to control the wavelength. An end mirror 2550 is included
to provide optical feedback and define an extended cavity. Minor
components may also be included, such as an array of micro-lenses
2515 or an array of apertures 2520
[0175] The manufacturing costs of laser arrays depends on component
count and on the cost to align and package the components.
Recently, the cost of manufacturing the VECSEL semiconductor die
has decreased and packing densities have increased as the growth
and fabrication of the semiconductor portion of a VECSEL has
improved. Studies by the inventors indicate that alignment and
packaging costs of fabricated devices are now a major cost factor
of a VECSEL. Conventional cavity designs require careful mechanical
or thermal alignment of individual laser elements. In particular,
conventionally several critical optical alignments are required.
The critical optical alignments require precise alignment to
achieve a high device efficiency and typically also require careful
design of packaging fixtures to maintain the critical optical
alignment over the lifetime of the laser. In contrast, in the
low-cost design embodiments of this invention, we aim to eliminate
the need to align each individual emitter of an array but instead
align all the array elements at once to maximize gain and nonlinear
conversion on all array emitters at once.
[0176] One aspect is that multiple emitters (gain elements) can be
arranged on a single die and attached to the same heat sink and
electrically wired in a single manufacturing step. By utilizing an
array of surface emitting lasers formed on a common die a large
number of lasers can be aligned and packaging at once, improving
manufacturability. Another advantage in using an array approach is
its increased reliability since the array does not have a
single-point-of-failure disadvantage of traditional lasers.
Possible degradation in power or even complete failure of some of
the array elements will not invoke unrecoverable degradation of the
entire light source and can be compensated for in ways that are not
available on traditional laser systems. In addition, the
reliability of nonlinear optical elements is greatly increased when
the power density is not too high as it could be in a
single-emitter laser of comparable power. Instead, the power
densities for each emitter can be kept at reasonably low
levels.
[0177] In one embodiment the end mirror is a segment with planar
facets to facilitate a simple alignment process between the end
mirror and the surface emitting laser array chip. The optical
design is correspondingly design to account for planar end mirrors.
For example, arrays of micro-lenses may be included to define the
mode waist. The other major optical components may also be
implemented as segments with planar facets to facilitate optical
alignment.
[0178] Another aspect of the present invention is that two or more
of the components in FIG. 25 may be combined into one optical
assembly (e.g., by monolithically bonding two or more components
together into one unit) to further reduce alignment complexity. In
some cases individual components provide multiple functions. In
other cases individual components are formed into larger optically
assemblies of components that are attached to each other, which is
facilitated by designing the components to have substantially
planar surfaces that may be attached to each other. Referring to
FIG. 26, the functionality of the wavelength control element and
end mirror may be combined into one unit by, for example, utilizing
a volume Bragg grating to provide optical feedback and frequency
control. Referring to FIG. 27, the functionality of the nonlinear
crystal, wavelength control element, and end mirror may be combined
by, for example, attaching a volume Bragg grating to a nonlinear
crystal. As illustrated in FIG. 28, the functionality of the
polarization control element, nonlinear crystal, wavelength control
element, and end mirror may be combined by, for example, attaching
a polarization control element to a nonlinear crystal which in turn
is mounted to a volume Bragg grating. As illustrated in FIG. 29,
the functionality of minor optical components may also be combined
with other functional blocks. For example, an array of micro-lenses
may be attached to a surface emitting laser array chip, such as an
array of dome lenses to help optimize the spatial mode of the laser
cavity and/or to focus light onto the nonlinear crystal. As another
example, the array of domes may further be coated to be
transmissive at the fundamental wavelength and highly reflective at
the second harmonic such that light at the second harmonic is
reflected to expand and have reduced spatial overlap a forward
propagating second-harmonic beams. As another example, an array of
apertures may be lithographically integrated with the surface
emitting laser array chip. As a consequence the number of optical
components that must be aligned and packaged in stable optical
alignment is reduced.
[0179] Realizing the advantages described above requires a number
of innovative design steps. These steps include but are not limited
to using volume Bragg gratings (new elements commercialized only
very recently), using pulsing of an array to maximize nonlinear
frequency conversion process with a simple and manufacturable
cavity design, using specially designed thin-film interference
filters, etc.
[0180] FIGS. 30 and 31 illustrate two main preferred cavity layout
embodiments. However extensions and modifications of these designs
such as folding the cavity, adding lens and aperture arrays for
mode control, and using other laser design options that keep the
design scalable are also within the scope of the present invention.
All of the designs shown illustrate array embodiments. It is
implied that these cavity designs can be equally used for
single-emitter configurations, one-dimensional arrays, and
two-dimensional arrays.
[0181] The cavity design shown in FIG. 30 comprises an
electrically-pumped, semiconductor, surface-emitting laser array
2105 and the extended cavity consisting of a polarizing
beamsplitter 2320, a nonlinear crystal 1615, and a volume Bragg
grating 2310. The design of semiconductor laser or laser array was
described in patents by Mooradian ("High power laser devices," U.S.
Pat. No. 6,243,407; "Efficiency high power laser device," U.S. Pat.
No. 6,404,797; "High power laser," U.S. Pat. No. 6,614,827;
"Coupled cavity high power semiconductor laser," U.S. Pat. No.
6,778,582) which are incorporated here by reference. A lens array
may be integrated with the chip or positioned separately in the
extended cavity. When integrated with the chip, the lens may be a
thermal lens generated by the heat created in the gain area or a
static lens, e.g. etched on the chip surface. Such a lens may help
to stabilize the spatial eigenmode of the laser and is especially
desirable when one wants to design the external cavity with a flat
output coupler for simplicity and/or cost reasons. As was noted
before in the discussion, the unique aspect of the surface-emitting
array configuration is its scalability--a large number of emitters
can be put on the same die which can be mounted on the same
heatsinks and fabricated without many extra steps that are required
for conventional discrete laser devices. The pitch between
neighboring surface emitters is optimized to minimize thermal
cross-talk. It was found experimentally and through simulations
that a pitch of about 2.5 and larger ratio to the gain aperture
diameter can be reached with minimum thermal cross talk and this
can provide packing densities accommodating many emitters in a
compact package. For example, one preferred design for a 3W
single-color laser source for projection display application can be
produced by a die of about 1 mm.times.7 mm cross section with 20-30
emitters of about 100-120 micron gain diameter. If necessary, the
arrangement of emitters on an array may be selected to be
non-periodic (with non-uniform pitch) to optimize thermal
performance and reduce the heating of emitters in the center of the
surface-emitting array chip.
[0182] Volume Bragg grating (VBG) is a wavelength-selective element
that is made of special glass with a periodic refractive index
variation written in it. Such an index variation can be designed to
produce a spectrally narrow high-reflectivity element that can help
to control the spectrum of the laser in a window selected by
design. While fiber Bragg gratings have been known for several
years in telecom laser design applications their volume
counterparts (VBGs) have been commercially available only recently.
The principles of such grating elements are described in U.S. Pat.
No. 6,586,141, "Process for production of high efficiency volume
diffractive elements in photo-thermal refractive glass" by O. M.
Efimov, L. B. Glebov, V. L. Smirnov, and L. Glebova, and U.S. Pat.
No. 6,673,497, "High efficiency volume diffractive elements in
photo-thermal refractive glass" by O. M. Efimov, L. B. Glebov, and
V. L. Smirnov. Previously, it was described how to use VBGs for
frequency stabilization of edge-emitting lasers and laser arrays
(G. Venus, V. Smirnov, L. Glebov, "Spectral Stabilization of Laser
Diodes by External Bragg Resonator", Proceedings of Solid State and
Diode Laser Technology Review, Albuquerque, N. Mex., June 2004, B.
L. Volodin, V. S. Ban, "Use of volume Bragg gratings for the
conditioning of laser emission characteristics," U.S. patent
application Ser. No. 10/884,524). Volume Bragg grating elements are
currently commercially available from Optigrate (Orlando, Fla.),
PD-LD (Pennington, N.J.), and Ondax (Monrovia, Calif.). Similar
holographic elements have also been fabricated and used in media
storage technology (e.g, by InPhase Technologies) and while we will
use the term volume Bragg grating (VBG) in the following
discussion, it is assumed that using such holographic grating
elements is also within the scope of this invention.
[0183] In the preferred embodiment, the nonlinear crystal is a
periodically poled nonlinear crystal such as periodically poled
KTP, LiNbO3, or LiTaO3. The advantage of using such crystals is
that they posses high nonlinearity and can be engineered for
nonlinear frequency conversion of the wavelength of choice.
Similarly, the surface-emitting semiconductor laser chip and volume
Bragg grating can be designed around the same wavelength.
Non-periodic (chirped) poling of crystals like KTP, LiNbO3, or
LiTaO3 to expand the nonlinear conversion bandwidth in temperature
and/or wavelength space. However, conventional bulk nonlinear
materials such as KTP, LBO, KNbO3, and others may also be used. In
some cases, it may be desirable to make optically contacted
assemblies of these crystals, for example, the thickness of
periodically poled materials may be increased by optically
contacting two, thinner pieces.
[0184] Finally, to define the linear polarization of the laser that
makes the nonlinear frequency conversion (second harmonic
generation) process efficient, the cavity needs an element
providing polarization control. The polarizing beamsplitter element
2320 shown in FIG. 30 provides polarization control function and
can also be used to re-direct the back-propagating
frequency-converted beam outside of the cavity. To perform these
functions, the two surfaces of such a polarizing beamsplitter need
to be coated to discriminate and provide high transmission for the
desired polarization at the fundamental laser wavelength and one of
the surfaces needs to be coated for high reflection at the
frequency converted wavelength. Alternately, the polarization may
be again controlled by the birefringence of nonlinear crystal as
described in U.S. patent application Ser. No. 10/734,553,
"Polarization control of an extended cavity laser" by A. V.
Shchegrov, A. Umbrasas, J. P. Watson, and D. Lee, the contents of
which are hereby incorporated by reference.
[0185] The design shows in FIG. 30 allows collecting both forward-
and backward-propagating frequency-converted beams for each
emitter. A turning mirror 3005 may be used to steer the light in
the same direction.
[0186] For example, semiconductor laser array can be designed for a
wavelength around 1064 nm--in this example, epitaxial structure can
be grown by such techniques as MOCVD or MBE on a GaAs wafer and
epitaxially grown mirrors and quantum wells can be targeted for
1064 nm. Volume Bragg grating can be designed to provide maximum
reflection at 1064 nm in a spectral bandwidth that is narrow enough
to stay within the phase-matching (or quasi-phase-matching)
bandwidth of the nonlinear material. The nonlinear material can be
properly selected to provide efficient second harmonic generation
from 1064 nm to 532 nm. Examples of such materials include PPKTP,
PPLN (which can be MgO-doped to increase its damage threshold),
PPLT, KTP, and others.
[0187] The cavity design of FIG. 30 would result in generating
second-harmonic beams propagating in two directions. While it is
possible to re-combine and overlap these beams collinearly by
reflecting the back-propagating beam by a dichroic coating on the
semiconductor chip or the nonlinear crystal, a simple collection of
both forward- and backward-propagating beams as shown in FIG. 30
may be acceptable in such applications as projection display light
sources. Additionally, a simple system avoids the complicated phase
control required to avoid destructive interference in a beam
combining scheme. However, overlapping the forward- and
backward-propagating beams by designing a proper dichroic coatings
preferably with a designed-in phase shift on reflection is also
within the scope of this invention. Note that the optical paths of
the forward and backward propagating second-harmonic beams may be
selected to be sufficiently long before recombination to reduce
interference effects. In particular, the forward and backward
propagating beams may be recombined after traversing optical path
lengths greater than their coherence length.
[0188] Other ways known in the art of laser design may be used to
improve the efficiency of second-harmonic light extraction. One way
to do this (not shown on the picture) is to fold the cavity into an
L-shape and return both the fundamental and the second-harmonic
light onto their forward paths. Such a design is also scalable to
array platform and is also within the scope of this invention.
[0189] Yet another scalable design that allows collinear
re-combining the forward- and the backward-propagating beams is
illustrated in FIG. 31. Here, a preferred low-cost embodiment of
monolithically bonded elements 3120 is shown but a design with
separate optical elements is also allowed. Only the second harmonic
beams are shown in dotted lines. Forward propagating second
harmonic beams are extracted through the volume Bragg grating.
Backward-propagating second harmonic beams have their polarization
rotated by ninety degrees (to avoid destructive interference and/or
dephasing effects) and are then reflected back to the volume Bragg
grating.
[0190] The embodiment shown in FIG. 31 uses a quarter-waveplate
3105 (for the second-harmonic wavelength) to rotate the
polarization of the second-harmonic light and then reflect it back
by a proper dichroic coating on the waveplate or on the
surface-emitting chip. This permits the backward-generated beams to
be recombined with the forward-propagating second-harmonic beams of
orthogonal polarization. This type of beam combining helps to avoid
potentially undesirable destructive interference between the beams.
In this embodiment, both polarizations of the second-harmonic light
will be collected from the resulting laser light source. A
waveplate that is quarter-wave for the second-harmonic, may be
designed to be half-wave for the fundamental wavelength so that it
does not affect the operation of the laser at the fundamental
wavelength. Waveplates of this type are commercially available.
Additionally, it is possible to sandwich the dichroic coating
between two identical waveplates that are rotated at 90 degrees to
each other. Such a system allows the transmitted light to receive
zero change in polarization while allowing the light reflected off
of the dichroic coating to be changed by an arbitrary degree, e.g.,
rotated by 90 degrees. Another advantage of this design is its low
cost packaging structure since several elements may be
monolithically bonded and only two subassemblies need to be
mechanically aligned to achieve optimum laser operation. As for the
thermal alignment which is usually necessary to achieve optimum
nonlinear conversion by the crystal, it is facilitated by the fact
that volume Bragg gratings are typically very stable with respect
to temperature variations with tuning rates of 0.1 nm/.degree. C.
or less. This means that only the nonlinear crystal needs to be
tuned to achieve the optimum conversion, e.g. by a low-cost
resistive heater, and other elements do not require similar active
temperature control. In the preferred embodiment, illustrated in
FIG. 3, only one element needs to be aligned for the whole array.
Furthermore, the alignment tolerances for this type of cavity
design are not very demanding. For example, for a gain diameter of
about 100 to 120 microns per emitter the tolerance for alignment
within about 5% of visible power is about one milliradian of
tilt.
[0191] Finally, both designs of FIGS. 30 and 31 are rather compact,
and, especially when the thermal lens in the semiconductor chip
stabilizes the cavity, only passive mechanical alignment steps may
be necessary to achieve the optimum laser operation.
[0192] The embodiment shown in FIG. 31 requires a separate
polarization control element. While it is still possible to use a
beamsplitter approach as shown in FIG. 30, it is preferable to keep
the design simple and integrate the polarization control function
with any of the existing optical elements (surface-emitting array,
waveplate, crystal, and VBG). One of the preferred approaches to do
that is to deposit a wire-grid polarizer on one of these elements
to reject the unwanted polarization and provide very low loss to
the desired polarization. Again, the idea of this approach is to
reduce the number of elements and alignment steps for low-cost,
high-volume manufacturing.
[0193] Both designs of FIGS. 30 and 31 and their extensions are
optimized for a low-cost, highly manufacturable platform of the
laser system. However, a low-cost design could make it more
challenging to achieve desired power target specifications in the
second harmonic conversion process. One way to increase the
efficiency of the second harmonic process is by designing the
cavity with a lens array or lens arrays that focusing the beams of
fundamental wavelength into beams with a small waists located in
the crystal. While this option is within the scope of this
invention, it is not the primary approach for achieving a low-cost
architecture with efficient frequency doubling. The preferred
approach for improving the efficiency of the second-harmonic
process is to drive the surface-emitting, electrically pumped laser
array with short pulses, e.g. 100 ns with large enough duty cycle,
e.g. 5-10% to increase the peak power in the fundamental-wavelength
intracavity beams and increase the average power in the
second-harmonic beams. Pulsed operation with high repetition rates
such as, e.g., 1 MHz can be as much acceptable in many applications
as continuous wave operation. For example, in display applications
such repetition rates will not be perceived by the human eye as a
pulsed operation and, therefore, such pulsed sources are acceptable
as much as continuous-wave sources. Furthermore, they can be even
desirable since pulsed operation can provide greater flexibility
for designers of display systems.
[0194] Another advantage of electrically pumped,
frequency-converted laser array sources is that these light sources
can be directly modulated at high rates such as, e.g. 25 MHz, by
modulating electrical drive. Modulation capabilities can be
desirable in such applications such as scanning-based laser
projection devices, etc.
[0195] Other designs which are within the scope of this invention
include laser arrays with frequency control elements other than
volume Bragg gratings. An example of such an element is a thin-film
coating interference filter, which can be design to provide
resonant narrow-bandwidth transmission at the fundamental design
wavelength.
[0196] One of the primary applications of embodiments of laser
light sources of the present invention is projection displays. For
these applications, it is desirable to have red, green, and blue
(RGB) colors to achieve a full color space representation for best
viewing experience. Embodiments described above are not specific to
any color and can be used to design a full RGB laser array light
source, which is a part of this invention. For example,
semiconductor gain material and mirror stacks can be designed
around 1260 nm, 1064 nm, and 920 nm, and by choosing proper poling
periods and optical coatings for nonlinear material such as lithium
niobate, and designing VBGs and beamsplitters or waveplates around
these wavelength, we obtain a laser array light source of 630 nm
(red), 532 nm (green), and 460 nm (blue).
[0197] The power in each color can be scaled by a simple increase
of the number of emitters in an array. In some cases, it is
desirable to use one-dimensional array of emitters. For example,
many periodically poled nonlinear crystals are fabricated with a
thin cross-section .about.0.5 mm which may make it difficult to fit
multiple rows of emitters along that dimension. In this case, it
may be preferable to scale the number of emitters in a single
dimension. Alternatively, a two-dimensional emitter array may be
used and when it is not desirable to increase the thickness of
nonlinear material, a subassembly of optically contacted thin
crystals may be used in place of a single crystal.
[0198] Applications of the present invention include light sources
for projection displays, illumination applications, automotive
lighting, and other consumer-electronics application. For example,
a projection display system based on coherent laser source may
suffer from speckle effect. However, the speckle effect is reduced
when the number of emitters is increased in a laser array. Further,
multi-longitudinal mode operation, which is the preferred
embodiment of this invention reduced coherence of each emitter and
thereby reduces speckle. This advantage is further promoted by the
pulsed operation of the laser. Yet another advantage is an
increased reliability of such an array because a failure in one
emitter does not mean a failure of the entire light source and
because the power density in nonlinear crystal and other optical
element is not as high as it would be in a single-emitting laser of
comparable power, e.g. at the multi-Watt level desirable for laser
light sources for rear-projection display televisions.
[0199] Lasers with these designs are compatible with methods of low
cost packaging. In particular, it is possible to make these lasers
compact and with simple alignments. In a preferred embodiment, the
critical alignment is between the surface of the laser array and
the surface of the output mirror, both of which are flat. The laser
is designed in such a way that the tolerances of this alignment can
be met passively, or at a minimum, passive alignment is sufficient
to achieve initial lasing, which can then be readily optimized.
This eliminates the need to search for initial operation. Such a
simple package could be readily incorporated into a multiple color
package, providing for further space reductions in a laser system
for display applications. An example of such packages is shown in
FIGS. 32A, 32B, and 32C. FIG. 32A shows a package for a single
array (e.g., an array for light of one color). FIG. 32B shows a
cross-section along line 3200. FIG. 32C shows a set of packaged
arrays, such as a set for generating red, green, and blue
light.
[0200] In the packages shown in FIGS. 32A, 32B, and 32C, there are
several elements that benefit the overall system. First, the
package does not require active temperature stabilization. This is
achieved through engineering of both the surface emitting laser, as
well as engineering of the package to have high thermal
conductivity and a small distance between the laser and the base of
the package. Second, the alignment can be performing using passive
alignment to high-precision fiducial marks of the laser package.
Systems such as those shown in FIGS. 32A, 32B, and 33C preferably
have heat dissipation equivalent to modern central processing unit
(CPU) chips, and so any of the efficient, low-cost cooling
solutions designed for these CPUs could be used for such a
laser.
[0201] Second, the system is designed with a minimum number of
elements. In the case of the units shown in FIGS. 32A and 32B,
there are only four distinct elements: the laser, the polarizer,
the nonlinear material, and the VBG. Furthermore, only the VBG
requires a critical alignment. Thus the packaging cost is reduced
due to the minimal number of critical alignments. Additionally the
reduction in component count also facilitates maintaining optical
alignment during the lifetime of the unit.
[0202] Third, the system is designed to be compact. In one
embodiment the package has a width less than 1.5 inches and a total
volume less than about two inches such that the packages lasers may
be utilized as a replacement for UHP lamps in a variety of
projection display systems. This is enabled by limiting the number
of elements, and by using arrays of lasers with simple, compact
cavities, instead of a single laser, which due to its higher
performance, will require more critical alignments, and
correspondingly more space and elements to make those
alignments.
V. VECSEL Design for Compact, High Efficiency, High Power Arrays to
Replace UHP Lamps in Display Systems
[0203] One application of laser arrays of the present invention is
as a replacement for conventional white light sources used in
projection displays. As previously described, a scalable,
manufacturable architecture permits the number of VECSELs to be
selected to achieve high power, reliable, substantially
speckle-free output. Additionally, experimental data indicated that
the VECSEL arrays can be extremely compact and have a high
efficiency. Experimental data demonstrates that individual VECSELs
may generate output powers in the visible on the order of 30 to 50
mW using pulsed-mode frequency doubling. Optimization of the pulse
width of pump-lasers has produced more than a factor of two
improvement in average pulsed SHG power. Thermal modeling has
demonstrated that for pulsed mode applications VECSELs can be
tightly packed with packing ratios on the order of 2:1 to 3:1.
[0204] The VECSEL gain element is preferably optimized for low
optical loses. With substrate thicknesses of 50 to 100 microns
using gain diameters of 100 to 400 microns the carrier gain
distributions are nearly top-hat for substrate doping levels down
to the mid E16 n-type.
[0205] It was discovered through empirical investigations that the
efficiency of VECSELs increases as the electrically pumped diameter
increases. Each gain element of a VECSEL may have current confined
to a pre-selected diameter using, for example, proton implants or
other techniques to limit the current injection to a preselected
diameter. Empirical studies demonstrated that arrays with a high
packing density may be utilized in pulse mode, decreasing the cost
of the semiconductor portion of the laser.
[0206] Referring to the FIG. 33, in a VECSEL with current injection
limited to a selected diameter there will be an electrically pumped
of a quantum well gain region. However, if the gain in the
electrically pumped region is sufficiently high and the diameter
sufficiently wide there will also be substantial amplified
spontaneous emission in the lateral plane of the quantum wells. As
a result, there will be an optically pumped annular region with low
optical losses as illustrated in FIG. 34. These effects are
particularly pronounced for high current pulsed mode operation and
large diameter VECSELs, such a VECSELs having an electrically
pumped diameter in the range of 80-150 microns. This annular
lateral pumping produces additional optical gain and a low loss
region. For large diameter devices that are highly pumped 40-60% of
the optical power can be in the optically pumped annular regions.
Another way to understand the effect of the annular lateral pumping
is that energy lost from lateral amplified spontaneous emission is
recovered as long as the external cavity mode is adjusted such that
the optical mode diameter recovers at least some of the power in
the annular region.
[0207] Empirical studies have demonstrated VECSELs with a 150
micron diameter electrically pumped gain region having a
circulating power in the extended cavity at the fundamental
wavelength approaching a few hundred watts. The high circulating
power density, in turn, improves the SHG output. Additionally, a
comparatively large electrically pumped diameter relaxes alignment
tolerances compared with narrow diameter VECSELs. Alternatively,
another way to understand large diameter VECSELs is that in
addition to additional benefits afforded by lateral optical
pumping, the large diameter produces lower optical coupling losses
for a particular degree of manufacturing misalignment.
[0208] The high efficiency of the VECSELs also reduces heat
dissipation, reducing cooling requirements. Additionally, the high
efficiency and high packing density permits a comparatively compact
module having a total volume of one or two cubic inches to generate
substantial power at several visible wavelengths, such as RGB.
[0209] FIG. 35 is chart comparing VECSELs (ECSEL column), LEDs, and
UHP lamps. VECSEL arrays fabricated in accordance with the
teachings of the present invention now have power, brightness,
efficiency, and cost characteristics superior to UHP lamps.
Additionally, the VECSEL arrays are extremely compact, eliminate
the need for color wheels and fans for DLP systems. In other
display applications, such as a 3LCD engine, the need for
additional polarizers, color filters, turning mirrors, and fly eye
lenses are eliminated. As consequence the VECSEL arrays are useful
as replacements for UHP lamps in many projection display
applications. Moreover, the compact size provides high coupling
into even comparatively small microdisplays. Thus the light sources
and methods of operation described in this patent applications
might be used in a variety of light engine architectures as
replacements to conventional LEDs and UHP lamps.
VI. Other Optimizations
[0210] Embodiments of the present invention may be practiced with a
variety of optimizations. Variations of the nonlinear crystal are
contemplated as being within the scope of the present invention.
The nonlinear crystal may, for example be periodically poled
materials such as periodically poled lithium niobate (PPLN,
periodically poled lithium tantalate (PPLT), periodically poled
potassium titanyl phosphate (PPKTP), or periodically poled rubidium
titanyl arsenate (PPRTA). The nonlinear crystal may be poled with a
chirped aperiodic pattern to increase the nonlinear conversion
spectral and temperature bandwidths. The nonlinear crystal may be a
bulk nonlinear material such as lithium triborate (LBO), potassium
titanyl phosphate (KTP), beta-barium borate (BBO), cesium lithium
borate (CLBO), or potassium niobate (KNbO3).
[0211] The gain element may be formed from a variety of
semiconductor materials, such as GaInAs, GaAlAs, GaInAsN, and
GaN.
[0212] The volume Bragg grating may be designed with a chirped
index of refraction pattern to provide reduce array coherence and
speckle.
[0213] Additional optical elements, such as fiber bragg gratings
may be included to provide frequency control.
VII. Combinations and Sub-Combinations
[0214] While the present invention has been described in regards to
various example, it will be understood that the various examples
may also be used in combinations and sub-combinations.
VIII. Other Applications
[0215] While the individual lasers and lasers arrays of the present
patent application have been described extensively in regards to
projection display applications, it will also be understood that
they may be utilized for other applications.
[0216] The foregoing description, for purposes of explanation, used
specific nomenclature to provide a thorough understanding of the
invention. However, it will be apparent to one skilled in the art
that specific details are not required in order to practice the
invention. Thus, the foregoing descriptions of specific embodiments
of the invention are presented for purposes of illustration and
description. They are not intended to be exhaustive or to limit the
invention to the precise forms disclosed; obviously, many
modifications and variations are possible in view of the above
teachings. The embodiments were chosen and described in order to
best explain the principles of the invention and its practical
applications, they thereby enable others skilled in the art to best
utilize the invention and various embodiments with various
modifications as are suited to the particular use contemplated. It
is intended that the following claims and their equivalents define
the scope of the invention
* * * * *