Interferometer for spatial chirp characterization

09964449 ยท 2018-05-08

Assignee

Inventors

Cpc classification

International classification

Abstract

Spectral interferometric systems and methods to characterize lateral and angular spatial chirp to optimize intensity localization in spatio-temporally focused ultrafast beams are described. Interference between two spatially sheared beams in an interferometer leads to straight fringes if the wavefronts are curved. To produce reference fringes, one arm relative to another is delayed in order to measure fringe rotation in the spatially resolved spectral interferogram. Utilizing Fourier analysis, frequency-resolved divergence is obtained. In another arrangement, one beam relative to the other is spatially flipped, which allows the frequency-dependent beamlet direction (angular spatial chirp) to be measured. Blocking one beam shows the spatial variation of the beamlet position with frequency (i.e., the lateral spatial chirp).

Claims

1. An optical system comprising: a first optical element or set of optical elements configured to receive a light beam having spatially chirped pulses of light and split the light beam into a first light beam and a second light beam; a second optical element or set of optical elements configured to receive the second light beam and provide an altered second light beam; a third optical element or set of optical elements configured to receive the first light beam and spatially flip the first light beam with respect to the altered second light beam thereby providing an altered first light beam; and an imaging spectrometer configured to generate an interferogram based on the altered second light beam and the altered first light beam.

2. The optical system of claim 1, further comprising a fourth optical element or set of optical elements configured to combine the altered first light beam and the altered second light beam into a third light beam, wherein the interferogram is based on the third light beam.

3. The optical system of claim 1, wherein the first optical element is configured to combine the altered first light beam and the altered second light beam into a third light beam, wherein the interferogram is based on the third light beam.

4. The optical system of claim 1, further comprising a fourth optical element or set of optical elements configured to receive the first light beam, provide a second altered first light beam, and vary a divergence of a fourth light beam resulting from the combination of the second altered first light beam and the altered second light beam.

5. The optical system of claim 4, wherein the fourth optical element or set of optical elements is configured to add a time delay to the first light beam.

6. The optical system of claim 4, wherein the fourth optical element or set of optical elements is mounted to a translational stage.

7. The optical system of claim 4, wherein the first light beam is received at either (i) the third optical element or set of optical elements or (ii) the fourth optical elements or set of optical elements at any one time.

8. The optical system of claim 7, wherein the imaging spectrometer is configured to generate a second interferogram based on combination of the second altered first light beam and the altered second light beam, wherein the second altered first light beam is received from the fourth optical element or set of optical elements.

9. The optical system of claim 8, wherein the second interferogram depicts one or more vertical fringe patterns.

10. The optical system of claim 1, wherein the first optical element or set of optical elements include a beam splitter, the second optical element or set of optical elements include a corner cube, and the third optical element or set of optical elements includes a prism.

11. The optical system of claim 10, further comprising a fourth optical element or set of optical elements configured to vary a divergence of a third light beam resulting from a combination of a second altered first light beam and the altered second light beam, wherein the fourth optical element or set of optical elements includes a corner cube.

12. The optical system of claim 11, further comprising a lens in an optical path of the third light beam, wherein the lens is located between the third optical element or set of optical elements and the imaging spectrometer.

13. An optical system for characterizing properties of a light beam, the system comprising: a first interferometer arrangement including a first optical element or set of optical elements, a second optical element or set of optical elements, and a third optical element or set of optical element, wherein, the first optical element or set of optical elements is configured to receive a light beam having spatially chirped pulses of light and split the light beam into a first light beam and a second light beam, the second optical element or set of optical elements is configured to receive the first light beam and add a delay in time to the first light beam thereby providing an altered first light beam, the third optical element or set of optical elements is configured to receive the second light beam and provide an altered second light beam; a second interferometer arrangement including the first optical element or set of optical elements, a second optical element or set of optical elements, and the third optical element or set of optical element, wherein, the first optical element or set of optical elements is configured to receive the light beam having spatially chirped pulses of light and split the light beam into the first light beam and the second light beam, the third optical element or set of optical elements is configured to receive the second light beam and provide the altered second light beam, and the second optical element or set of optical elements is configured to receive the first light beam and spatially flip the first light beam with respect to the altered second light beam thereby providing a second altered first light beam; and an imaging spectrometer configured to generate an interferogram based on the altered second light beam and the second altered first light beam, wherein a measurement of spatial chirp based on an interference fringe pattern in the interferogram is generated.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The present disclosure is described in conjunction with the appended figures:

(2) FIG. 1 is a schematic diagram depicting an optical system in accordance with embodiments of the present disclosure;

(3) FIG. 2 is a schematic diagram depicting additional details of an optical system in a accordance with embodiments of the present disclosure;

(4) FIG. 3 depicts a rotation of a fringe pattern resulting from beam divergence in accordance with embodiments of the present disclosure;

(5) FIG. 4 is a chart depicting the inverse radius of curvature versus change in lens position z (points) and a fit line providing slope to extract a focal length of a lens in accordance with embodiments of the present disclosure;

(6) FIG. 5 is a schematic diagram depicting additional details of an optical system in a accordance with embodiments of the present disclosure;

(7) FIG. 6A depicts a Spatial Fourier transform of an interferometer arrangement of FIG. 5 with zero time delay where the sloping lines represent the variation of beam angle with frequency in accordance with embodiments of the present disclosure;

(8) FIG. 6B depicts a variation of the angular chirp rate with angular adjustment of the second grating in a double-pass compressor (points) where the fit (line) may be used to determine the position wherein the gratings are parallel and the x axis, listed as change in grating angle, is for an arbitrary angle; accordingly, the zero angle does not correspond to the zero angle between gratings;

(9) FIG. 7A depicts a lateral spatial chirp from an SSTF compressor in accordance with embodiments of the present disclosure;

(10) FIG. 7B depicts a lens conversion from lateral to angular spatial chirp with magnification to see such fringe curvature in accordance with embodiments of the present disclosure; and

(11) FIG. 8 depicts a method in accordance with embodiments of the present disclosure.

DETAILED DESCRIPTION

(12) The ensuing description provides embodiments only, and is not intended to limit the scope, applicability, or configuration of the claims. Rather, the ensuing description will provide those skilled in the art with an enabling description for implementing the described embodiments. It being understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the appended claims.

(13) In accordance with at least one embodiment of the present disclosure, an optical system 100 to characterize beams that are spatially chirped at the input, to an experiment for example, so that spatial chirp can either be eliminated or introduced in a controlled manner is described. That is, by introducing spatial chirp, each beamlet will share the same spatial characteristics, allowing much of the analysis to be treated analytically. For simplicity, a Gaussian form for each beamlet is utilized. The beam in the spatio-spectral domain may be written as Equation 1.

(14) E ( r , ) = E 0 ( ) exp ( i k .Math. r ) exp ( - r 2 w 2 ) exp ( i c r 2 2 R ) Equation 1

(15) Here, E.sub.0()=A()exp(i()) is the complex spectral envelope, where () is the overall spectral phase that might be controlled by the compressor or propagation through optical material. The angular chirp is encoded in the phase k.Math.r=(/c)(x sin(.sub.x)+z cos(.sub.x)), where .sub.x() describes the frequency-dependent angle. The coordinate representing the location of the maximum intensity of the beamlets, r.sub.=[(xx.sub.s).sup.2+y.sup.2].sup.1/2, is shifted in the x direction by x.sub.s(). This lateral spatial chirp can be expressed either in terms of the lateral or the angular chirp rates, or , respectively: x.sub.s()=(.sub.0) or x.sub.s()=f tan(.sub.x()) with .sub.x()=(.sub.0). The local spot size w(z,) evolves with propagation as a Gaussian beam. It is important to note that z=0 represents the plate at which the beamlets cross and that the position(s) of the beamlet waists, in general, can be at some other z position. The last term in Equation 1 describes the evolution of the local radius of curvature R(z,), which, in general, could be frequency-dependent. The location of the R= plane coincides with that of the beamlet waist.

(16) Because a uniform reference beam may not be available or may be difficult to create, two spectral interferometry (SI) techniques that are self-referenced are utilized. To measure the divergence, a spatial shear combined with spatially resolved SI is utilized. A prior approach consisted of adjusting the overlap, relative angle, and relative timing of the beams. However, in such an approach with zero time delay and a fixed crossing angle, precise crossing-angle calibration was required. In accordance with embodiments of the present disclosure, optics are utilized to avoid a crossing angle and time delay is utilized to provide the reference fringes, thereby eliminating the calibration step. Accordingly, the beam is not analyzed after the focusing optic, which can introduce aberrations. However, such an approach is well-suited to characterizing any corrections that might be introduced before focusing.

(17) Angular chirp can be detected by adding an additional reflection in one interferometer arm to combine beams with relative spatial inversion. This inversion technique was used to perform nonlinear autocorrelation to measure PFT. A local measurement of the PFT does not provide sufficient information to predict the profile elsewhere: PFT cannot only arise from angular chirp .sub.x() but may also arise from a combination of lateral chirp x.sub.s() and overall spectral phase (). In accordance with embodiments of the present disclosure, a linear measurement in the spectral domain measures angular chirp directly. In contrast with prior interferometric techniques, a Fourier inversion is utilized to obtain the frequency-dependent angular distribution. Moreover, as will be shown, the beamlet divergence affects the result of the angular chirp measurement because it combines spatial and angular shear.

(18) In accordance with embodiments of the present disclosure, an optical system 100 for characterizing divergence and spatial chirp is provided. The optical system 100 may include a beam source 104, which generally provides an input light 108 as an input to a first optical element or a set of optical elements 106. The input light 108 may be a light beam and may be a spatially chirped light beam. The first optical element or set of optical elements 106 conditions the input light 104 and provides first and second output light 148 and 152 for analysis and characterization. In some embodiments, the first optical element or the first set of optical elements 106 may include one or more beam splitters 112, an optical element or set of optical elements 116 specific to divergence measurement, and an optical element or a set of optical elements 136 specific to analyzing angular and transverse chirp. The optical element, set of optical elements 116, and/or a first interferometric arrangement may comprise a top arm including a corner cube 120 and a right arm including a corner cube 128. The optical element, a set of optical elements 136, and/or a second interferometric arrangement may comprise a top arm including a prism, such as a roof prism 140, or a pair of mirrors and a right arm including a corner cube 128. The corner cube 128 may be shared; alternatively, or in addition, each right arm may include a different corner cube 128.

(19) A beam with a spatial chirp evolves with propagation (e.g., angular chirp leads to lateral chirp). Any characterization must be performed at a well-defined plane to predict the form of the beam elsewhere in the system. In accordance with embodiments of the present disclosure, the path length to the measurement plane is the same for both configurations 116 and 136. Pulse characterization can be performed either before introducing spatial chirp or at a plane where all the frequency components are spatially overlapped. Autocorrelation has been performed at a spatio-temporal focus; single-pulse options include multiphoton intrapulse interference phase scan and recording the spectrum of the second harmonic as the chirp of the pulse is varied.

(20) Although other techniques have been demonstrated for making spatial chirp measurements, techniques such as GRENOUILLE measures first-order lateral spatial chirp and pulse front tilt but does not characterize spatial wavefront. Having a well-characterized reference beam allows for the measurement of arbitrary spatio-temporal couplings. Multishot scanning examples are fiber-based SI, gated angular spectrum, and spatially resolved SPIDER. Examples of single-shot referenced measurements are STRIPED-FISH and spatially resolved SI.

(21) As previously mentioned, the two interferometric arrangements utilized are shown schematically in FIG. 1. The first interferometric arrangement has two corner cubes to produce two output beams with lateral (x) and temporal shear () and is used to measure the beam divergence. The second interferometric arrangement replaces one of the corner cubes with an optical element, such as a roof prism 140, to spatially invert one beam, so that, if 0, spatial interference of the spectral components can occur to characterize the angular chirp. In both cases, the output is directed into an imaging spectrometer 124, which samples the beams across a line aty=0. Such an imaging spectrometer 124 may include an imaging element to perform such sampling. Relative time delay is utilized to allow the phase to be extracted with the Fourier analysis, as described below. Complete spatio-spectral characterization starts with measuring divergence utilizing the first interferometer arrangement and setting to zero, if desired. The second interferometer arrangement most accurately measures angular chirp. Note that lateral chirp can be converted to angular chirp by placing a long focal length lens one focal length in front of the entrance slit of the imaging spectrometer 124. Further processing of the resulting interferogram may be performed at the processing device 156 having processor 160 and memory 164. Alternatively, or in addition, the processing device 156, processor 160, and/or memory 164, alone or in combination, may be incorporated into the imaging spectrometer 124.

(22) A detailed schematic of the first interferometric arrangement is illustrated in FIG. 2. As previously mentioned, a beam source 204 generally provides an input light 208A as an input to a first optical element or a set of optical elements 106 comprising the first interferometric arrangement. The first interferometric arrangement includes two corner cubes 120 and 128 to produce two output beams with lateral (x) and temporal shear () and is used to measure the beam divergence. That is, the input light 208A is split at the beam splitter 112 resulting in input light 208B and input light 208D. Input light 208B is directed to the top arm including the corner cube 120 and results in conditioned input light 208C. Input light 208D is directed to the right arm including the corner cube 128 and resulted in conditioned input light 208E. Conditioned input light 208C and conditioned input light 208E may then be combined or otherwise made to overlap, using the beam splitter 112 and/or another optical element, such that the conditioned input light 208C and conditioned input light 208E are directed into the imaging spectrometer 124.

(23) Measurement of divergence occurs using the first interferometric arrangement, such as the interferometric arrangement depicted in FIG. 2. Such an arrangement, is a spectrally resolved variation of a shearing interferometer to provide a divergence interferogram 132. The glass shear plate or the air-edge interferometers, which are not spectrally resolved, have limited application for ultrashort pulses. With spectral resolution, the temporal offset () may be used to produce reference fringes in x space. As a side note, a variant of the cyclic or Sagnac shearing interferometer has been previously utilized in such a configuration. The variant of the cyclic or Sagnac shearing interferometer can be adjusted to produce spatial and temporal shear; however, the corner-cube design allows decoupled adjustment of both shears and an easy change to the spatially inverted configuration of the second interferometric arrangement. The output intensity measured at the spectrometer 124 is provided according to Equation 2.
I(x,)=|E(x+x,)exp(i)+E(xx,)exp(i)|.sup.2Equation 2

(24) The full spatial and temporal shifts are 2x and 22, respectively. The interferogram 132 is insensitive to the overall spectral phase () and the angular chirp .sub.x(). When Equation 1 and Equation 2 are combined and reduced, the interferogram becomes Equation 3.

(25) I ( x , ) = .Math. A ( x , ) .Math. 2 ( 2 + e i 2 ( 1 x cR x + ) + e - i 2 ( 1 x cR x + ) ) Equation 3

(26) The addition of time delay provides reference fringes so that the fringes rotate in x space as the divergence is changed. As further illustrated in FIG. 3, a converging beam will rotate the fringes counterclockwise (e.g., interferogram 132A), a diverging beam clockwise (e.g., interferogram 132C), and vertical fringes indicate a collimated beam (e.g. interferogram 132B).

(27) As a means to test the divergence measurement, a beam from a Ti:sapphire oscillator was passed through a spatial filter and recollimated with a nominal f=200 mm lens placed on a translation stage to vary the beam divergence. The fringe contrast decreases where the beams were not well overlapped spatially.

(28) Rather than tracking the fringes for direct fringe rotation measurement, process the interferogram 132 was processed using Fourier analysis. The image is inverse-Fourier transformed in the spectral direction. An AC peak is selected with a mask and re-centered on the grid. Fourier-transforming back to the spectral domain, the complex second term in Equation 3 is obtained. For each frequency component in the spectrum, the phase was fitted to a line in the x-direction and the slope for each frequency component was divided by the local value of 2/c to yield the quantity sin .sub.c, where .sub.c is the local angle between wavefronts. The two crossing wavefronts depend on the beam radius of curvature and the spatial shear between the beams, sin .sub.c=x/R. The shear axis then measured by placing a lens in front of the interferometer to focus the input beam to the entrance slit. The distance between the two resulting spots is equal to 2x. From the angle between the wavefronts and the spatial shear, R is calculated, which in general, is a function of ; in the present case, however, all spectral components were averaged for a better signal-to-noise ratio.

(29) Such a measurement was made for several positions (z) of the collimating lens and the data points were plotted with a Gaussian propagation fit (see FIG. 4) according to a measured distance between the lens and spectrometer slit of 132 cm, a focal beam radius of 25 m, and the fitted focal length of the lens is 18.95 cm. The percent difference between this is measured then the focal length is extracted and the focal length is calculated, according to the lens maker's equation (18.01 cm), which results in 5.2%. As previously discussed, further processing of the interferogram may be performed at the processing device 156 having processor 160 and memory 164. Alternatively, or in addition, the processing device 156, processor 160, and/or memory 164, alone or in combination, may be incorporated into the imaging spectrometer 124.

(30) A detailed schematic of the second interferometric arrangement is illustrated in FIG. 5. As previously mentioned, a beam source 204 generally provides an input light 508A as an input to a first optical element or a set of optical elements 106 comprising the second interferometric arrangement. The second interferometric arrangement includes a prism or pair of mirrors, such as roof prism 140, in the top arm and a corner cube 128, or a triplet of mirror, in the right arm, yielding an uneven number of bounces in the two arms and therefore spatially flipping the beams relative to each other in the vertical direction. That is, the input light 508A is split at the beam splitter 112 resulting in input light 508B and input light 508D. Input light 508B is directed to the top arm including the roof prism 140 for example, and results in conditioned input light 508C. Input light 508D is directed to the right arm including the corner cube 128 and resulted in conditioned input light 508E. Conditioned input light 508C and conditioned input light 508E may then be combined or otherwise made to overlap, using the beam splitter 112 and/or another optical element, such that the conditioned input light 508C and conditioned input light 508E are directed into the imaging spectrometer 124.

(31) For the measurement of spatial chirp with the second interferometer arrangement, the beams are inverted along the x-direction without spatial shear:
I(x,)=|E(x,)exp(i)+E(x,)exp(i)|.sup.2Equation 4

(32) When there is no lateral spatial chirp, the resulting spatial chirp interferogram (e.g. 144) is insensitive to both () and R(). This results in an interferogram 144 that is similar to Equation 3, but instead of the constant local crossing angle sin .sub.c=x/R, there is the frequency dependent angle, sin .sub.x=().

(33) Three tests varying the amount of spatial chirp were performed; a large angular chirp from a single transmission grating, a small spatial chirp from a detuned CPA grating compressor, and a large lateral spatial chirp from a SSTF single-pass compressor. For the first test of the angular chirp, a 110 grooves/mm transmission diffraction grating was imaged with unit magnification through the interferometer to the spectrometer entrance slit, thereby overlapping all spectral components there. The fringe curvature seen in the interferogram 144 results from the nonlinear term (quadratic for a linear chirp rate) in the interferogram 144. Fourier transforming the interferogram 144 in each direction results in a central DC peak and two crossed lines as depicted in FIG. 6A. Transforming in the spectral direction provides an angled line that represents the pulse front tilt I(x,). Transforming instead in the spatial direction as shown gives a line representing the angular chirp I(f.sub.x,), where f.sub.x=(sin .sub.x)/. To use the Fourier processing outlined above, time delay was added between the beams to fully separate the interference terms from the central amplitude term. The delay should satisfy (.sub.eff(x)=2/c sin(.sub.x)x+) to achieve sufficient separation. For such a measurement of the angular chirp with the transmission grating, a value of 203 rad/nm was obtained, which is in excellent agreement with the expected value of 205 rad/nm. As previously discussed, further processing of the interferogram may be performed at the processing device 156 having processor 160 and memory 164. Alternatively, or in addition, the processing device 156, processor 160, and/or memory 164, alone or in combination, may be incorporated into the imaging spectrometer 124.

(34) To test the sensitivity of the angular chirp measurement, the output of a double-pass grating compressor was directed into the interferometer. The angular chirp for several angular settings of the second diffraction grating (in combination with the retroflection roof mirror in the compressor) was then measured. FIG. 6B shows the measured angular chirp rate as a function of this grating angle, as shown in Equation 5.

(35) d d = - 2 d tan [ d ] cos [ i ] Equation 5

(36) .sub.2 is the movement of grating 2 from parallelism to grating 1, d is the diffracted angle off the second grating, and .sub.d is the incident angle on the first grating. The data can be fitted to this line to find the optimum grating angle for zero spatial chirp. The grating compressor was set to that angle position and a final interferogram was obtained to confirm zero chirp. The result was an extremely small 0.014 rad/nm of angular spatial chirp. As depicted, one of the points in FIG. 6B resides off the fitted line: for this point, the slope is close to zero, and the variations in the curve are consistent with imperfections found in the /4 per in. surface figure of the interferometer optics (leading to variations of about 10 rad/nm). Better surface figure optics as well as interferometric grade tolerance of the roof and corner reflectors will increase the precision and accuracy of the measurement.

(37) The final test was to characterize a single-pass grating compressor used to prepare the beam for SSTF. Two 1200 grooves/mm gratings were placed approximately 130 mm apart. A strong lateral spatial chirp in this case is seen in the variation of the central position of the intensity with wavelength. The lateral chirp was measured without interference by finding the spatial centroid of intensity versus wavelength shown in FIG. 7A. The lateral spatial chirp was transformed to angular chirp by placing a lens 148 in front of the interferometer. With the spectrometer entrance slit at the focal plane, the lateral to angular chirp conversion with a lens can be expressed as Equation 6.

(38) d d = 1 f d x 1 d Equation 6

(39) With regard to Equation 6, f is the focal length of the lens, d/d is the angular chirp rate (), and dx.sub.1/d is the lateral chirp rate (). If the input beam to the single-pass compressor is collimated, the spectral components will overlap with no tilt to the intensity envelope in the (,x) camera plane. The angular chirp can be measured by analyzing the interferogram, as shown in FIG. 7B. A spectrometer grating groove density higher than the one used (300 groove/mm) would allow a larger time delay to be used, giving better separation of the modulation peak from the zero-frequency peak. Both the extracted lateral chirp rate (220; 576 mm/nm) and the lateral chirp rate converted from measured angular chirp rate (220; 565 mm/nm) compare well with the grating compressor analytical model spatial chirp rate (220; 451 mm/nm).

(40) In the general case, where there is input divergence, there is a frequency-dependent spatial shear, which couples the measured wavefront angles and the lateral spatial chirp. In this case, the local angle versus frequency plots for spatially chirped beam have the following slope: m=2(/R+). In the simplest case, the input beam can be set to have no divergence so the lateral spatial chirp in this limit does not matter and a pure angular spatial chirp can be obtained. Or, instead, the divergence using the first interferometer arrangement can be measured and the combination of lateral and angular chirp from the slope can be extracted. Furthermore, the crossing plane of the beamlets can be forced to be at the beam waist with the following condition on the radius of curvature in relation to the angular and lateral chirp rates, 1/R=/.

(41) Referring now to FIG. 8, a method 800 to characterize spatial chirp will be described in accordance with embodiments of the present disclosure. Method 800 generally beings at step S804, where a light beam is provided to an optical system, such as optical system 100 previously described. Such light beam may be received at step S808, where the light beam encounters a beam splitter, such as beam splitter 112, resulting in first and second light beams. Divergence may then be adjusted at step S812 utilizing a first interferometric arrangement, such as the first interferometric arrangement described in FIGS. 1 and 2. Step S816 may be an optional step. At step S820, the first light beam is spatially flipped with respect to the second light. For example, a second interferometric arrangement, such as the second interferometric arrangement described in FIGS. 1 and 5 may be utilized. The light beams may then be combined at step S824 and one or more fringe patterns may be generated at step S828 utilizing an imaging spectrometer 124 for example. The one or more fringe patterns may be generated in a manner as previously described. At step S832, spatial chirp may be determined based on the one or more fringe patterns as previously described. For instance, the interferogram may be Fourier transformed and spatial chirp may be determined as previously described. Method 800 may end at step S840.

(42) As provided herein, a series of interferometric measurements that can determine several parameters (angular and lateral spatial chirp, beam divergence), which are required to characterize and align spatially chirped optical systems have been described. In principle, the spectrally resolved spatial shear measurement may provide higher-order wavefront information, such as coma or spherical aberration. Further, the measurements in several setups have been tested and are in good agreement with predicted results. Such a system generally finds use in the alignment of spatially chirped ultrafast systems as well as in more conventional chirped pulse amplification compressors.

(43) Specific details were given in the description to provide a thorough understanding of the embodiments. However, it will be understood by one of ordinary skill in the art that the embodiments may be practiced without these specific details.

(44) While illustrative embodiments of the disclosure have been described in detail herein, it is to be understood that the inventive concepts may be otherwise variously embodied and employed, and that the appended claims are intended to be construed to include such variations, except as limited by the prior art.