Holography is a technique for recording and reconstructing the wavefront produced by an object, or for synthesizing a wavefront that would be produced by a specified object. In its classical optical form, it uses interference between an object wave and a reference wave to encode information about both the amplitude and phase of light. When appropriately illuminated, the resulting record, called a hologram, redirects light by diffraction so that an observer or an imaging system receives a reconstructed wavefront. This distinguishes holography from ordinary photography, which directly records the spatial distribution of light intensity rather than the complete information needed to reconstruct a wavefront.
Holography is important both as an imaging method and as a means of measuring and manipulating waves. It can reproduce depth and changing perspectives, detect extremely small deformations, measure transparent specimens without staining, and form optical components without conventional curved surfaces. Its principles also apply to electron waves, X-rays, and sound. Modern holography includes photographic recording, numerical reconstruction of camera data, and computer-generated holograms designed for programmable optical devices. These methods share a concern with wavefront information, but differ substantially in how that information is acquired, stored, and reconstructed.
Origins and historical development
The conceptual foundation of holography is the distinction between a wave’s intensity and its phase. Optical detectors usually respond to intensity averaged over many oscillations; they do not directly follow the rapidly varying electric field of visible light. During the nineteenth and early twentieth centuries, interference and diffraction established ways to infer phase relationships indirectly. Holography combined these effects into a two-stage procedure: encode a wavefront through interference, then reconstruct it through diffraction.
Dennis Gabor developed the method in 1947 while seeking to improve the resolving power of the electron microscope, and published his proposal in 1948. In his original arrangement, illumination transmitted through a specimen served partly as the reference, while light scattered by the specimen supplied the object wave. This geometry later became known as in-line holography. Gabor demonstrated the principle with visible light, but the available sources had limited coherence, and the reconstructed image was contaminated by an overlapping conjugate image, usually called the twin image. These difficulties restricted early practical use.
The invention of the laser in 1960 transformed the field. A laser could supply intense light with sufficiently stable phase relationships over useful distances. In the early 1960s, Emmett Leith and Juris Upatnieks at the University of Michigan developed off-axis holography, drawing on concepts from communications and radar. By tilting the reference beam relative to the object beam, they separated the desired reconstruction from unwanted diffraction components. Their demonstrations included three-dimensional objects whose reconstructed appearance changed as the observer moved.
Independently, Yuri Denisyuk developed reflection holography, reporting the approach in 1962. His method connected wavefront recording with the wavelength-selective behavior of thick interference structures. Reflection holograms could eventually be viewed under suitable white-light illumination, greatly extending holography beyond laboratory arrangements requiring laser reconstruction. Stephen Benton’s rainbow hologram, introduced in 1969, provided another route to white-light viewing by sacrificing vertical parallax while retaining horizontal perspective.
Gabor received the Nobel Prize in Physics in 1971 for the invention and development of the holographic method. Subsequent advances shifted the field from exclusively photographic techniques toward electronic sensing and computation. Computer-generated holograms were already being investigated in the 1960s, while the 1994 work of Ulf Schnars and Werner Jüptner helped establish direct camera recording followed by numerical reconstruction as a practical form of digital holography. Later improvements in image sensors, computing, lasers, and programmable optical devices expanded applications in microscopy, manufacturing, optical trapping, and display research.
Interference recording and wavefront reconstruction
At a recording plane, a monochromatic scalar object field can be represented by the complex quantity
where is its amplitude and its phase. A reference field , whose form is known or controlled, illuminates the same plane. The measured intensity is proportional to
where an asterisk denotes complex conjugation. The first two terms contain the separate beam intensities. The cross terms contain their relative phase and therefore encode information that an ordinary intensity image of the object alone would not retain.
This recording process requires adequate optical coherence. Temporal coherence governs how far the optical path lengths of the two beams may differ while still producing stable fringes. Spatial coherence governs phase relationships across the illuminated region. The beams must also have compatible polarizations, because mutually orthogonal polarization components do not produce the required interference term. Motion or vibration during an exposure can shift the fringes and reduce their contrast. The necessary stability is set by the wavelength and the geometry, not merely by the apparent sharpness of the object.
For a simplified thin hologram whose amplitude transmission varies linearly with the recorded intensity, illuminating the developed hologram with the reference field produces several outgoing components. One is proportional to , reproducing the object field apart from a scale factor. Another contains a conjugated object field and produces the conjugate image. Additional light forms an undiffracted or zero-order background. Actual materials may instead modulate phase, and their responses may be nonlinear, but the underlying separation into useful and unwanted components remains central to hologram design.
A reconstruction may form a virtual image, from which the light appears to diverge, or a real image, toward which the reconstructed light converges. The image is not simply painted onto the hologram surface. Its apparent position follows the curvature and direction of the emerging wavefront. Consequently, an observer may focus at different depths and see different perspectives while moving within the available viewing region.
The familiar claim that every fragment of a hologram contains the whole image requires qualification. When light from each object point reaches much of the recording aperture, a small fragment can indeed reconstruct a broad view of the scene. Cutting the aperture, however, reduces collected light, angular coverage, and spatial resolution, and may remove perspectives hidden by occlusion. The object’s information is distributed across the recording, but a fragment does not retain all the information or image quality of the complete hologram.
Recording geometries and major hologram types
Holograms are classified along several independent dimensions. In-line and off-axis describe the relationship between object and reference waves; transmission and reflection describe how recording and reconstruction light travel through the medium; thin and volume describe how the medium’s thickness affects diffraction. A single hologram can therefore be an off-axis, transmission, volume hologram. These categories are not interchangeable, and each addresses a different physical or practical property.
In in-line holography, the reference and object fields propagate in nearly the same direction. A common arrangement illuminates small particles or a sparse specimen with a plane or spherical wave. Unscattered illumination supplies the reference, and scattered light carries the object information. This arrangement is compact and uses the detector’s spatial bandwidth efficiently, but the object reconstruction, its conjugate, and the background tend to overlap. Sparse specimens, multiple recording distances, and iterative reconstruction can help separate these contributions. In-line geometry is especially useful in lensless particle imaging.
Off-axis holography introduces an angular tilt between the reference and object beams. Their interference produces a spatial carrier: fine fringes whose frequency depends on the angle between the beams. For two plane waves meeting at an angle in air, the fringe spacing is
where is the wavelength. A larger angle improves separation of diffraction orders, but creates finer fringes that demand greater recording resolution. In a digital system, this separation also consumes part of the camera’s available spatial-frequency range.
Transmission holograms are normally viewed with illumination arriving from the side opposite the observer. Reflection holograms return the reconstructed light toward the illumination side. Denisyuk recording places the incident reference and the object-returned wave on opposite propagation paths through the recording layer, creating fringes with a strong depth-dependent structure. Thick reflection holograms can select a narrow wavelength range from white light, allowing reconstruction without a monochromatic viewing source.
Rainbow holograms provide white-light transmission viewing by restricting one direction of the recorded angular information, usually with a slit in an intermediate recording arrangement. Horizontal movement reveals changing perspective, whereas vertical movement predominantly changes the visible color. They therefore do not preserve full parallax. Holographic stereograms take another approach: multiple perspective images, captured photographically or rendered by computer, are encoded into small recording elements. The resulting display can reproduce directional views of a scene without requiring that a single optical exposure directly record the entire original object wavefront.
Recording materials, volume effects, and replication
A recording material must resolve the interference fringes and convert their exposure pattern into a lasting modulation of optical transmission, phase, or both. Ordinary photographic resolution is not always sufficient: holographic fringes can have spacings substantially below one micrometre. Material choice determines exposure sensitivity, attainable diffraction efficiency, scattering, environmental stability, and whether the recording can be replicated or rewritten.
Silver-halide emulsions have been widely used because they combine high spatial resolution with useful sensitivity. Their light-sensitive grains produce a developed image, and bleaching processes can convert much of the resulting absorption pattern into a phase-modulating structure. Dichromated gelatin can produce efficient phase holograms through exposure-dependent changes in gelatin followed by processing. Its performance depends strongly on humidity, processing, and protection of the finished layer. Photopolymers offer another important route: illumination initiates polymerization and associated material redistribution, producing a patterned refractive index. Some formulations require relatively little wet processing.
The distinction between thin and volume holograms concerns how the reconstructed wave interacts with the recorded structure. A thin hologram is often approximated as a two-dimensional modulation imposed on a passing wave. A volume hologram contains a three-dimensional fringe structure extending through a thickness large enough for selective, coupled diffraction to become important. Its behavior is related to Bragg diffraction: strong reconstruction occurs only when illumination has a suitable wavelength and direction relative to the recorded fringe planes.
This selectivity has major consequences. A thick reflection hologram can diffract a narrow spectral band from white light, while a volume transmission grating can couple light efficiently into a selected outgoing direction. Multiple holograms can be superimposed in the same volume using different reference angles or wavelengths and later addressed separately. The available refractive-index modulation is finite, however, so multiplexed recordings share the material’s dynamic range. Thickness, absorption, shrinkage, and departures from the intended fringe geometry also affect efficiency and cross-talk.
Mass-produced security images commonly use surface-relief structures rather than individually exposed thick recording layers. An original pattern is transferred through mastering and replication processes into microscopic relief, often embossed into a polymer and combined with a reflective coating. The relief controls the phase of reflected or transmitted light. Not every embossed diffractive image reconstructs a full object wavefront, although such products are often grouped commercially under the name “hologram.”
Processing changes can visibly alter a hologram. Shrinkage changes fringe spacing and orientation, shifting the wavelength or angle at which reconstruction is strongest. In display work this can produce an unintended color change; in precision optical elements it can move the device away from its design condition. Material characterization is therefore part of optical design, rather than merely a finishing step after exposure.
Digital recording and computer-generated holography
Digital holography replaces a photographic recording plate with an electronic image sensor and reconstructs the field numerically. A camera still measures intensity rather than optical phase directly, so a reference wave, multiple exposures, or additional constraints are generally needed to recover the complex object field. Once recovered, that field can be propagated computationally to different distances, corrected for known aberrations, or converted into quantitative phase and amplitude images.
In off-axis digital holography, the tilted reference shifts the cross terms away from the central background in spatial-frequency space. A Fourier transform allows one cross term to be isolated with a numerical filter. After shifting it to the origin and transforming back, the object field can be recovered up to reference-dependent factors. The method can operate from a single exposure, which is valuable for moving specimens. Its principal trade-off is that separation between terms reduces the sensor bandwidth available for the object image.
Phase-shifting holography instead records several interferograms with controlled reference-phase offsets. Combining the measurements separates the background from the complex interference term without requiring a large off-axis carrier. This can use the detector’s spatial bandwidth more efficiently, but sequential exposures are sensitive to specimen motion and phase drift. Simultaneous phase-shifting arrangements distribute the measurements across channels or sensor regions, trading temporal constraints for optical complexity or spatial sampling.
Numerical propagation may use the Fresnel approximation, convolution methods, or the angular-spectrum method. In the latter, a measured field is decomposed into plane-wave components. Under a homogeneous-medium scalar model, propagation over distance is represented by
where , and and are transverse spatial frequencies expressed as wave numbers. Discrete sampling, finite detector size, and the treatment of evanescent components constrain the calculation.
Computer-generated holography reverses the acquisition problem: a desired optical field is specified, and a diffractive pattern is calculated to produce it. The pattern may be fabricated permanently or displayed on a spatial light modulator. Because many devices control mainly phase or mainly amplitude, an arbitrary complex field cannot simply be copied onto them. Encoding schemes and iterative optimization distribute the desired information among available degrees of freedom. Applications include beam shaping, optical trapping, aberration correction, microscopy, and experimental three-dimensional displays. These calculated holograms need not originate from an actual object or an interference exposure.
Holographic metrology and quantitative microscopy
Holographic interferometry compares wavefronts corresponding to different states of an object. A hologram may record an object before loading, heating, or vibration, and its reconstructed field may then interfere with light from the changed object. Alternatively, two states can be recorded together, or their complex fields can be compared digitally. The resulting fringes reveal changes in optical path rather than merely changes in visible surface appearance.
For a displacement vector , the phase change is governed by the projection of displacement onto a sensitivity vector determined by illumination and observation directions. In the simple case of normal-incidence reflection from a surface displaced along its normal,
A phase change of therefore corresponds to a displacement of . This does not mean that every instrument measures displacement to that accuracy: attainable measurement uncertainty also depends on noise, geometry, surface behavior, environmental stability, and calibration.
These techniques have been used to study deformation of mechanical components, thermal expansion, vibration modes, and subsurface defects that alter a surface’s response to stress or heating. Time-averaged holography records many vibration cycles and produces characteristic fringe patterns related to vibration amplitude. Pulsed illumination can freeze rapid events, while digital systems can retrieve phase maps across sequences of frames. Rough surfaces scatter light into granular interference patterns called speckle, which both enable wavefront comparison and introduce decorrelation when the surface moves too far or changes its scattering properties.
Digital holographic microscopy extends phase measurement to transparent specimens. Under a simple projection approximation, the phase delay relative to a surrounding medium is
where is the specimen’s refractive index and is that of the medium. A single measurement therefore combines thickness and refractive-index contrast; it does not independently determine both. Measurements at multiple angles, with appropriate scattering models, can support three-dimensional refractive-index reconstruction.
Phase maps are normally obtained modulo . Recovering a continuous optical-path map may require phase unwrapping, which becomes difficult near discontinuities, low-signal regions, or rapidly varying structures. Multiple wavelengths can extend the unambiguous measurement range, but introduce additional demands on wavelength knowledge and dispersion modeling. In biological imaging, holographic methods can measure cell shape, motion, and optical-path changes without staining. Interpretation of such measurements requires a physical model; a phase image alone does not uniquely identify molecular composition or establish a medical diagnosis.
Electron, X-ray, acoustic, and other wave holographies
Holography is not restricted to visible light. Its essential requirement is access to a wave field, together with a means of encoding or recovering the field’s phase relative to a reference. The implementation changes greatly across the electromagnetic spectrum and for material waves, because sources, detectors, interactions with matter, and available optical components differ. In some fields the term includes numerical wavefront reconstruction rather than a directly viewable photographic hologram.
Electron holography uses the wave properties of the electron, described by quantum mechanics. In a common off-axis transmission-electron-microscope arrangement, an electron biprism overlaps a wave that has passed through a specimen with a reference wave traveling through a nearby region. Their interference fringes encode relative phase. Reconstruction can reveal phase shifts caused by the specimen’s mean inner potential, electrostatic fields, and magnetic vector potential.
Interpreting an electron hologram requires separating these contributions rather than treating phase as a direct map of a single material property. Specimen thickness, diffraction within crystals, charging, and instrument aberrations can all matter. Controlled changes of specimen orientation, magnetization, or applied conditions may help distinguish different contributions. The method is particularly useful because electric and magnetic structures can modify electron phase even when their effects are difficult to identify in a conventional intensity image.
X-ray holography faces different constraints. Short wavelengths permit high spatial resolution, but sufficiently coherent illumination and suitable reference structures are technically demanding. In Fourier-transform holography, a specimen and a small reference aperture are illuminated together, and their far-field diffraction pattern is recorded. Transforming the intensity pattern produces separated cross-correlation terms that contain an image of the specimen. Reference size, specimen-reference separation, detector sampling, and source coherence constrain the result. The method is related to, but distinct from, other forms of coherent diffraction imaging that recover phase through iterative constraints without a separately fabricated reference.
Acoustic holography reconstructs sound fields from measurements over a surface. Microphone arrays can measure complex sound pressure relative to a timing reference, and numerical propagation can infer fields elsewhere. Near-field acoustic holography can recover detailed source information by measuring close enough to retain rapidly decaying spatial components. Back-propagating those components amplifies noise, making regularization important.
Microwave and radio-frequency holographic methods likewise exploit coherent measurements, often using electronically accessible phase information. Applications include antenna characterization, scattering measurements, and wavefront control. Their common mathematical basis does not make their instruments interchangeable: the relevant wavelengths, near-field distances, detector responses, and interactions with materials determine what each implementation can resolve.
Optical elements, security devices, and data storage
A holographic optical element uses a recorded or calculated diffraction structure to perform an optical function. Instead of forming an image of an object for direct viewing, it may focus light, redirect a beam, separate wavelengths, or couple light into a waveguide. Recording the interference of suitable wavefronts creates a structure that can transform one into the other during reconstruction. A spherical and a plane wave, for example, can produce an element with lens-like behavior.
Such elements can be thin and lightweight, and several functions can sometimes be combined within one recording medium. Their response is often strongly wavelength- and angle-dependent, especially for volume structures. This dependence can be useful in a narrowband filter or display combiner but problematic when broad spectral or angular acceptance is required. Holographic elements used in near-eye optical systems must balance efficiency, viewing range, color behavior, transparency, and unwanted diffraction. They do not automatically reproduce the broadband performance of a conventional refractive lens.
Security applications exploit images and optical changes that are difficult to reproduce with ordinary printing or photography. Embossed diffractive features appear on banknotes, identity documents, payment cards, and product packaging. A feature may change color, switch between images, or reveal different information as the illumination or viewing angle changes. More specialized devices can combine overt visual effects with microscopic patterns or machine-readable structures. “Holographic” security products include a broad range of diffractive technologies, not all of which preserve a complete three-dimensional wavefront.
The security value of such a feature is not a physical guarantee against copying. It depends on the complexity and control of origination and replication, the specificity of the observable effect, and its integration with other document or packaging features. A superficially similar rainbow effect may be easy to imitate even when reproducing the intended fine structure is much harder. Consequently, visual appeal and authentication performance are different design objectives.
Holographic data storage records information as two-dimensional pages within a photosensitive volume. A spatial light modulator encodes a page onto an object beam, which interferes with a reference beam inside the medium. Different pages can be multiplexed by changing the reference angle, wavelength, position, or another addressing parameter. Reading reconstructs a page onto a detector array, offering parallel retrieval rather than strictly bit-by-bit access.
Practical storage systems face limits from material dynamic range, cross-talk, shrinkage, readout geometry, detector noise, and error correction. Recording more pages into the same volume generally reduces the signal available from each page unless material performance also improves. High laboratory storage density therefore does not by itself establish a reliable storage product: long-term stability, repeatable manufacturing, access speed, and compatibility with recording and readout hardware are separate requirements.
Three-dimensional displays and fundamental limitations
A holographic display seeks to deliver wavefronts consistent with a three-dimensional scene. Ideally, moving an eye changes the visible perspective, binocular vision supplies appropriate disparity, and accommodation responds to the apparent distance of objects. These properties distinguish wavefront reconstruction from a flat image that merely contains perspective cues. They also distinguish it from many stage illusions and projection effects popularly advertised as holograms.
Pepper’s ghost, for example, uses partial reflection to combine a scene with an image located elsewhere; it does not reconstruct the optical wavefront of a three-dimensional object through a holographic fringe pattern. Similarly, a rotating light display can form a luminous image within a swept volume without using holographic encoding. Such devices may provide compelling depth effects, but their physical mechanisms and viewing constraints differ from holography.
Electronic holographic displays face a severe sampling problem. A sampled optical modulator can generate only a limited range of diffraction angles. For pixel pitch , a useful approximate bound on the representable transverse spatial frequencies gives
under simple on-axis sampling assumptions. At a wavelength of nanometres and a pixel pitch of micrometres, this corresponds to a half-angle of about . Encoding choices, device structure, and optical arrangements modify the practical field, but the underlying relationship remains: wider viewing angles require finer spatial modulation.
A large display with a wide viewing zone therefore needs an exceptionally large number of controllable optical degrees of freedom. Pixel count alone is not sufficient; pixels must also have suitable phase response, fill factor, speed, and optical efficiency. Full-color operation adds wavelength-dependent diffraction and calibration requirements. Real-time scenes require rapid calculation of the holographic pattern, while available computing speed, data transfer, and device refresh rate constrain image complexity and motion.
Coherent illumination introduces speckle and other interference artifacts. Temporal averaging, wavelength diversity, angular diversity, and optimized phase patterns can reduce them, but may trade away efficiency, sharpness, or refresh capacity. Imperfect modulation can produce an undiffracted bright spot and unwanted diffraction orders. Viewing-zone expansion and eye-tracked systems can direct information more selectively, but involve additional optical components or tracking requirements.
Ultimately, holography remains subject to finite aperture, diffraction, sampling, and noise. It cannot reconstruct spatial information that was never captured or encoded, nor can an ordinary far-field optical hologram reproduce arbitrarily fine evanescent detail. A complete assessment of a display must therefore distinguish reconstructed depth from screen size, viewing angle from eye-box size, and nominal pixel count from the number of independently controlled optical modes.