Regions of the Electromagnetic Spectrum

A chemist who wants to know what molecule is sitting in a flask has no way to look inside it directly. Instead, she shines light of a chosen energy at the sample. She then asks which photons survive, which disappear, and which are re-emitted at a different color. The answer depends entirely on which region of the electromagnetic spectrum she chooses. This is because different photon energies interrogate different pieces of molecular and atomic architecture. These include the tight inner-shell electrons of an X-ray photon, the valence electrons of an ultraviolet photon, the bond vibrations of an infrared photon, and the nuclear spins of a radiofrequency photon. This matters because spectrochemical analysis is built entirely on matching the analytical question to the correct spectral window. Asking "what functional groups are present" calls for infrared radiation. Asking "how much iron is in this water sample" calls for ultraviolet-visible or atomic spectroscopy. Asking "what is the oxidation state of this manganese atom in a mineral" may call for X-ray absorption near an edge. Choosing the wrong region is not a subtle error. It is asking a question the photons in hand are physically incapable of answering. This is because the energy simply does not match any transition available to the sample. The big picture is that the electromagnetic spectrum is a continuous ladder of photon energies. But molecules and atoms have discrete, quantized energy levels. So only in narrow regions does light have the right energy to drive a particular class of transition. This builds on the wave and photon relationships that describe Properties of Electromagnetic Radiation — wavelength, frequency, and photon energy. This article maps the spectrum region by region. It covers which transition each region drives, and what kind of information it yields. It also lists the realistic wavelength/energy numbers a working analyst should have memorized. This article delivers a working atlas of the spectrum. It gives quantitative boundaries for the UV, visible, IR, and X-ray regions. It identifies the physical transition (electronic, vibrational, core-electron) that dominates in each. It includes worked numerical examples converting between wavelength and energy at realistic values. It also flags the traps that catch students who assume "more energetic" always means "more useful."

Key takeaways

The spectrum spans roughly 15 orders of magnitude in photon energy; analytically useful chemistry lives mostly between X-ray and radiofrequency.
X-rays (≈ 0.01–10 nm, keV energies) eject or excite core (inner-shell) electrons — used in XRF and XAS for elemental and oxidation-state analysis.
UV-visible light (≈ 190–800 nm, ~1.5–6.5 eV) promotes valence electrons between molecular orbitals — the basis of UV-Vis absorption and atomic emission/absorption.
Infrared light (≈ 2.5–50 μm, ~0.025–0.5 eV) drives molecular vibrations (bond stretches and bends) — the fingerprint region for functional groups.
Microwave and radiofrequency photons carry far less energy and probe molecular rotation and nuclear/electron spin states (rotational spectroscopy, NMR, EPR).
The photon energy $E = hc/\lambda$ sets a hard lower bound on what transition a given wavelength can drive — no amount of intensity substitutes for the right energy.
Key idea: each region of the spectrum is defined by which quantized energy gap it matches, so the analytical question dictates the region, not the reverse.

Where this lives today

Every modern analytical laboratory routes samples to instruments organized precisely along these spectral boundaries. A pharmaceutical quality-control lab uses UV-Vis spectrophotometry (200–800 nm) to assay drug concentration via characteristic chromophore absorbance, while a synthetic chemistry lab down the hall uses FTIR (2.5–25 μm) to confirm a newly synthesized compound's functional groups by their vibrational fingerprint — the same molecule, interrogated in two spectral regions that answer entirely different structural questions. Atomic absorption and inductively coupled plasma optical emission spectrometry (ICP-OES), both operating in the UV-visible region, remain the workhorse techniques for trace-metal determination in environmental water testing, clinical toxicology (blood lead, mercury), and food safety screening. X-ray fluorescence spectrometers, exploiting core-electron transitions in the keV range, are now handheld field instruments used by geologists, recyclers sorting scrap alloys by elemental composition, and museums performing nondestructive pigment analysis on paintings — a Fe Kα line at 6.40 keV or a Pb Lα line at 10.55 keV identifies an element without ever touching the artifact. Synchrotron-based X-ray absorption spectroscopy (XAS/EXAFS) pushes this further, resolving not just elemental identity but oxidation state and local coordination geometry, essential in catalysis research and battery-materials development where the local environment of a transition-metal center governs performance. At the low-energy end, rotational (microwave) spectroscopy identifies interstellar molecules in radio astronomy by their exquisitely narrow rotational line spacings, while nuclear magnetic resonance — operating on nuclear spin transitions in the radiofrequency range under a strong applied magnetic field — remains the single most information-rich technique in organic structure elucidation and, in its imaging form (MRI), a cornerstone of clinical medicine. Building outward from this atlas of regions, the next natural steps are to examine Energy Levels and Molecular Transitions in more quantitative depth and then to see how a single region — the UV/visible/near-IR — is engineered into an instrument in Introduction to Atomic Absorption Spectroscopy.

Ask the tutor: If a compound's UV-Vis spectrum shows $\lambda_\text{max} = 640$ nm, what does that tell you about its HOMO–LUMO gap in eV, and why would adding more conjugation shift this value?

The physical picture

Electromagnetic radiation is conventionally organized by wavelength $\lambda$ , frequency $\nu$ , or photon energy $E$ , related by

E = h\nu = \frac{hc}{\lambda}

with $h = 6.626\times10^{-34}\ \text{J s}$ (Planck's constant) and $c = 2.998\times10^{8}\ \text{m/s}$ . Because $E$ and $\lambda$ are inversely related, short-wavelength photons (X-rays, UV) are high-energy and interact with tightly-bound electrons, while long-wavelength photons (IR, microwave, radio) are low-energy and interact with the loosely-spaced levels of vibration, rotation, and spin. A molecule or atom has a ladder of quantized energy levels — electronic, vibrational, rotational, and nuclear-spin — each ladder with a characteristically different rung spacing. Core-electron levels are separated by keV; valence-electron (molecular orbital) levels by 1–10 eV; vibrational levels by tenths of an eV; rotational levels by meV or less; and nuclear-spin levels (in a magnetic field) by µeV. A photon can only drive a transition if its energy matches (resonates with) the gap between two levels the molecule actually possesses. This is why the spectrum is naturally divided into regions defined by which class of transition dominates there, not by arbitrary human convention alone — the divisions track real physics. The diagram below lays the regions out on a log-wavelength axis together with the characteristic transition each region drives.

Figure

The electromagnetic spectrum, from X-ray to radiofrequency, with the transition class dominant in each region

Within the visible band, wavelength maps to perceived color, and this mapping matters directly for UV-Vis spectrophotometry: the color an absorbing solution appears is the complement of the color it absorbs. A solution that absorbs at 620 nm (orange) appears blue, because blue light is transmitted while orange is removed.

Figure

Visible spectrum with complementary color relationships used in UV-Vis spectrophotometry

The core ideas

The single equation governing everything in this map is the photon energy relation already introduced:

E = h\nu = \frac{hc}{\lambda}

where:

$E$ = energy of one photon (J, or converted to eV for chemical convenience),
$h$ = Planck's constant, $6.626\times10^{-34}\ \text{J s}$ ,
$\nu$ = frequency (Hz, s $^{-1}$ ),
$c$ = speed of light in vacuum, $2.998\times10^{8}\ \text{m/s}$ ,
$\lambda$ = wavelength (m, converted to nm, μm, etc. as convenient).

Reading: for a given photon, wavelength and energy trade off exactly — halve the wavelength and you double the energy. This single inverse relationship is the entire reason the spectrum divides into functionally distinct regions: each factor-of-10 change in $\lambda$ corresponds to a factor-of-10 change in $E$ , sweeping past entirely different classes of quantized energy gaps. A second essential relation converts photon energy into the more chemically intuitive unit of energy per mole of photons, since spectroscopic transitions are often discussed in kJ/mol or eV per molecule:

E_\text{mol} = N_A h \nu = \frac{N_A hc}{\lambda}

where:

$E_\text{mol}$ = energy per mole of photons (J/mol, often reported in kJ/mol),
$N_A$ = Avogadro's number, $6.022\times10^{23}\ \text{mol}^{-1}$ ,
other symbols as above.

Reading: this equation is what lets you compare, on the same footing, the ~5 eV (≈ 480 kJ/mol) needed to break some chemical bonds against the ~3 eV (≈ 290 kJ/mol) carried by a typical near-UV photon at 400 nm — explaining, among other things, why UV radiation but not visible or IR radiation is capable of photochemically breaking bonds (as in ozone photolysis or sunburn).

Derivation APhoton energy at a representative wavelength in each region

Goal. Show explicitly, with the equation $E = hc/\lambda$ , the numerical energy carried by a photon typical of the X-ray, UV, visible, and infrared regions, and confirm the orders of magnitude claimed in the overview.

Step 1 — Set up the constants and a convenient eV form. Rather than repeatedly multiplying by $6.626\times10^{-34}$ and $2.998\times10^{8}$ and then dividing by $1.602\times10^{-19}$ to convert joules to electron-volts, it is efficient to precompute the constant $hc$ in eV·nm:

hc = (6.626\times10^{-34}\ \text{J s})(2.998\times10^{8}\ \text{m/s}) = 1.9865\times10^{-25}\ \text{J m}

Converting to eV·nm ( $1\ \text{eV} = 1.602\times10^{-19}\ \text{J}$ , and $1\ \text{m} = 10^{9}\ \text{nm}$ ):

hc = \frac{1.9865\times10^{-25}\ \text{J m}}{1.602\times10^{-19}\ \text{J/eV}} \times 10^{9}\ \text{nm/m} = 1240\ \text{eV}\,\cdot\,\text{nm}

so for any wavelength in nanometers, $E(\text{eV}) = 1240/\lambda(\text{nm})$ .

Step 2 — X-ray photon, $\lambda = 0.154$ nm (the Cu Kα line used in X-ray fluorescence and diffraction).

E = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{0.154\ \text{nm}} = 8.05\times10^{3}\ \text{eV} = 8.05\ \text{keV}

This is squarely in the keV range needed to eject a core (1s) electron from mid-row transition metals — the physical basis of X-ray fluorescence (XRF) elemental analysis.

Step 3 — Ultraviolet photon, $\lambda = 254$ nm (a common UV-C germicidal/analytical line).

E = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{254\ \text{nm}} = 4.88\ \text{eV}

Converting to a molar quantity:

E_\text{mol} = (4.88\ \text{eV})(96.485\ \text{kJ mol}^{-1}\text{eV}^{-1}) = 471\ \text{kJ/mol}

This is comparable to many single-bond dissociation energies (C–C ≈ 348 kJ/mol, C–H ≈ 413 kJ/mol), which is exactly why 254 nm UV light drives both valence electronic transitions and photochemistry.

Step 4 — Visible photon, $\lambda = 520$ nm (green light, near the middle of the visible band).

E = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{520\ \text{nm}} = 2.38\ \text{eV} \quad\Rightarrow\quad E_\text{mol} = 230\ \text{kJ/mol}

This is enough to promote a valence (HOMO→LUMO type) electron in a conjugated chromophore but is well below most single-bond dissociation energies — consistent with visible light being spectroscopically useful without being generally photochemically destructive.

Step 5 — Infrared photon, $\lambda = 5.80\ \mu\text{m}$ (a typical C=O stretch region, ≈ 1724 cm $^{-1}$ ). Converting micrometers to nanometers, $\lambda = 5800$ nm:

E = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{5800\ \text{nm}} = 0.214\ \text{eV} \quad\Rightarrow\quad E_\text{mol} = 20.6\ \text{kJ/mol}

This is roughly two orders of magnitude below a bond dissociation energy, matching the far smaller spacing of vibrational levels compared to electronic levels — IR photons can only nudge a bond between vibrational quantum states, never break it.

Step 6 — Compare orders of magnitude. Collecting results: X-ray ≈ 8050 eV, UV ≈ 4.9 eV, visible ≈ 2.4 eV, IR ≈ 0.21 eV. Each region differs from its neighbor by roughly one to three orders of magnitude, confirming the qualitative picture that X-ray, UV/visible, and IR photons address fundamentally different classes of quantized transitions (core-electron, valence-electron, and vibrational, respectively).

Go deeper · Derivation B · 1 minConverting a spectroscopic transition's energy gap into the wavelength region where it must be probed

Derivation BConverting a spectroscopic transition's energy gap into the wavelength region where it must be probed

Goal. Given a molecule's known HOMO–LUMO energy gap, predict the wavelength at which it will absorb, and confirm this places the transition in the correct spectral region — the inverse problem to the one above, and the one an analyst actually faces when interpreting an unknown spectrum.

Step 1 — State the resonance condition. Absorption occurs when photon energy equals the energy gap between two states:

E_\text{photon} = \Delta E_{\text{gap}}

Using $E_\text{photon} = hc/\lambda$ , solve for the absorbing wavelength:

\lambda = \frac{hc}{\Delta E_{\text{gap}}}

Step 2 — Insert a realistic gap. Take a moderately conjugated organic chromophore (analogous to a substituted azobenzene or a cinnamate ester) with $\Delta E_\text{gap} = 3.54\ \text{eV}$ , a value obtained independently from a computational or empirical estimate of its π→π* transition.

Step 3 — Solve using the $hc = 1240$ eV·nm shortcut.

\lambda = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{3.54\ \text{eV}} = 350\ \text{nm}

Step 4 — Check the region assignment. 350 nm falls within the near-UV band (roughly 200–400 nm), consistent with the chromophore being a conjugated π system: extended conjugation lowers $\Delta E_\text{gap}$ and pushes $\lambda_\text{max}$ toward, and eventually into, the visible region — precisely the mechanism by which highly conjugated dyes absorb visible light and appear colored, while simple, unconjugated alkenes (larger gap, ~ 6–7 eV) absorb only in the vacuum UV and appear colorless.

Step 5 — Limiting case. As $\Delta E_\text{gap} \to 0$ (an infinitely conjugated system, formally the free-electron limit), $\lambda \to \infty$ : the absorption redshifts without bound, which is exactly the trend observed in homologous series of polyenes and cyanine dyes, where each added conjugated unit measurably redshifts $\lambda_\text{max}$ by tens of nanometers, eventually moving the transition from UV through visible into the near-IR.

Worked examples

Example 1 (foundational) — Converting a stated IR absorption from wavenumbers to wavelength and energy. An infrared spectrum reports a carbonyl stretch at $\tilde{\nu} = 1715\ \text{cm}^{-1}$ . Find the corresponding wavelength in micrometers and the photon energy in both eV and kJ/mol. Wavenumber $\tilde{\nu}$ is defined as $1/\lambda$ with $\lambda$ in cm, so:

\lambda = \frac{1}{\tilde{\nu}} = \frac{1}{1715\ \text{cm}^{-1}} = 5.831\times10^{-4}\ \text{cm} = 5.831\ \mu\text{m}

Converting to nanometers, $\lambda = 5831$ nm, and using $E = 1240\ \text{eV}\,\cdot\,\text{nm}/\lambda$ :

E = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{5831\ \text{nm}} = 0.2127\ \text{eV}

Converting to a molar energy:

E_\text{mol} = (0.2127\ \text{eV})(96.485\ \text{kJ mol}^{-1}\text{eV}^{-1}) = 20.5\ \text{kJ/mol}

This value — 5.83 μm, 0.213 eV, 20.5 kJ/mol — sits comfortably in the mid-infrared "fingerprint-adjacent" carbonyl region (typical C=O stretches fall 1650–1750 cm $^{-1}$ , i.e., 5.7–6.1 μm), confirming the assignment is self-consistent across all three units. As a limiting-case check: if the wavenumber were instead halved to 857.5 cm $^{-1}$ (deeper into the fingerprint region, characteristic of C–Cl or skeletal bending modes), the wavelength would double to 11.66 μm and the energy would halve to 0.106 eV — the linear inverse scaling between $\tilde\nu$ and $E$ is preserved exactly, as required since both are directly proportional to $1/\lambda$ ... actually since $E\propto\tilde\nu$ directly (not $1/\lambda$ inverted twice), halving $\tilde\nu$ halves $E$ directly, which is what was found.

Example 2 (applied) — Selecting the correct spectral region for an atomic absorption determination of calcium. A water-treatment laboratory needs to quantify trace calcium (as Ca) in a groundwater sample by atomic absorption spectroscopy. The Ca resonance line lies at 422.7 nm. An instrument technician proposes instead using an infrared source at 4.5 μm because "IR is gentler and won't photobleach the sample." Evaluate whether this substitution is physically possible, using photon energies. The Ca 422.7 nm transition is a valence electronic transition (4s → 4p) with an energy gap of:

\Delta E = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{422.7\ \text{nm}} = 2.934\ \text{eV}

An IR photon at 4.5 μm (4500 nm) carries:

E_\text{IR} = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{4500\ \text{nm}} = 0.2756\ \text{eV}

The ratio of required energy to available IR photon energy is:

\frac{\Delta E}{E_\text{IR}} = \frac{2.934\ \text{eV}}{0.2756\ \text{eV}} = 10.6

A single IR photon carries roughly one-tenth the energy needed to drive the calcium 4s→4p transition. Since atomic absorption is a single-photon resonance process (the atom absorbs one photon per transition, not ten sequential lower-energy photons combining coherently under ordinary lamp intensities), the IR substitution is not merely "gentler" — it is physically incapable of exciting this transition at all. The technician's proposal must be rejected: only light in the near-UV/visible region, matched in energy to $\Delta E \approx 2.93$ eV, can populate the excited state that atomic absorption spectroscopy measures. This example generalizes the point made in the overview: region selection is dictated by the transition's energy gap, not by convenience or instrument availability.

Example 3 (exam/competition-level) — Distinguishing an XRF edge energy from a valence UV transition in a mixed mineral sample. A geochemist examines an unknown mineral with an X-ray fluorescence spectrometer and observes an emission line at $\lambda = 0.1937$ nm, tentatively assigned to the Fe Kα line (literature value 0.1937 nm, corresponding to a transition energy of 6.40 keV). A second, much weaker feature appears in a separate diffuse-reflectance UV-Vis scan at $\lambda = 970$ nm, in the near-infrared edge of the instrument's range. The geochemist wonders whether the 970 nm feature could actually be a weak second-order (harmonic) reflection of the same Fe Kα X-ray, which would invalidate its assignment as an electronic transition. First, confirm the Fe Kα assignment using $E = hc/\lambda$ :

E_{K\alpha} = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{0.1937\ \text{nm}} = 6402\ \text{eV} = 6.40\ \text{keV}

This matches the accepted Fe Kα energy (6.40 keV) to three significant figures, confirming the core-level (1s → 2p) assignment. Now test whether the 970 nm feature could plausibly be related to this X-ray line by a simple harmonic (integer wavelength multiple) relationship, as sometimes occurs with diffraction-grating stray light:

\frac{\lambda_{970}}{\lambda_{K\alpha}} = \frac{970\ \text{nm}}{0.1937\ \text{nm}} = 5008

For a genuine second-order diffraction artifact, the ratio would need to be a small integer (2, 3, 4 …), reflecting light of the same source appearing at $n\lambda$ in a diffraction pattern. A ratio of 5008 is not a small integer relation and is instead simply the ratio of two entirely different spectral regions: 970 nm corresponds to

E_{970} = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{970\ \text{nm}} = 1.278\ \text{eV}

which is a typical energy for a $d$ – $d$ ligand-field transition or a spin-forbidden transition in an Fe(II)/Fe(III) mineral — an entirely distinct valence-electron phenomenon from the 6.40 keV core-level fluorescence. The two features arise from unrelated physical processes (core-electron ejection versus $d$ -orbital crystal-field splitting) despite appearing in the same sample, and the huge disparity in photon energy (6402 eV vs 1.28 eV, a factor of ≈ 5000) is itself the proof: no plausible harmonic or instrumental artifact bridges five spectral decades. The geochemist should report both features independently — Fe Kα confirms the presence of iron; the 970 nm feature independently constrains its oxidation state and coordination environment.

Limiting cases, scope & common traps

Region	Typical λ range	Typical photon energy	Transition probed	Common analytical use
X-ray	0.01–10 nm	0.1–100 keV	core (inner-shell) electron ejection/excitation	XRF, XRD, XAS/EXAFS
Vacuum/far UV	10–200 nm	6–120 eV	valence electrons, requires evacuated optics (air absorbs)	specialized electronic spectroscopy
Near UV	200–400 nm	3.1–6.2 eV	valence π→π, n→π transitions	UV-Vis absorption, fluorescence
Visible	400–800 nm	1.55–3.1 eV	valence transitions in conjugated/colored species	colorimetry, Vis absorption
Near/mid IR	2.5–50 μm	0.025–0.5 eV	molecular vibration (stretch, bend)	FTIR functional-group ID
Far IR/microwave	50 μm–1 mm+	<0.025 eV	rotational transitions	rotational spectroscopy
Radiofrequency	mm–m, with applied B field	μeV range	nuclear/electron spin flip	NMR, EPR

The chemist behind it

The systematic division of the electromagnetic spectrum by wavelength owes its foundations to William Herschel, who in 1800 discovered infrared radiation while measuring the temperature rise produced by different colors of sunlight dispersed through a prism — and noticed the warmest reading fell just beyond the red end, where no visible light fell at all. This simple thermometer experiment first demonstrated that "light" as the eye perceives it is only a narrow slice of a much larger radiative continuum, a discovery that seeded the entire discipline of spectrochemical analysis by regions.

Origins · 1 minFriedrich Wilhelm (William) Herschel and how this was discovered

Friedrich Wilhelm (William) Herschel

1738–1822 · Georgian-era England, astronomer and experimental natural philosopher

Herschel, a Hanoverian-born musician turned astronomer in England, is best known for discovering the planet Uranus in 1781, but his 1800 experiment on solar heat proved equally consequential for chemistry. He passed sunlight through a prism, placed blackened-bulb thermometers at each visible color, and — testing for a control — placed one just past the red band where no color was visible. That "blank" thermometer registered the highest temperature of all, revealing radiant heat beyond the red: infrared radiation. Herschel's methodical use of a control thermometer, careful attention to reproducibility, and willingness to trust an anomalous null-region reading over intuition exemplify the same experimental rigor that underlies modern spectrochemical method development.

History

- 1800 — William Herschel discovers infrared radiation using dispersed sunlight and thermometers placed beyond the red band.

- 1801 — Johann Wilhelm Ritter discovers ultraviolet radiation by observing that invisible light beyond violet darkens silver chloride paper faster than visible violet light.

- 1859 — Gustav Kirchhoff and Robert Bunsen establish that each element produces a unique atomic emission/absorption line spectrum, founding quantitative atomic spectroscopy.

- 1895 — Wilhelm Röntgen discovers X-rays, opening core-electron spectroscopy and diffraction-based structural analysis.

- 1900s–1920s — Planck, Einstein, and Bohr establish the quantum relation

E=h\nu

and quantized atomic energy levels, providing the theoretical basis for why each spectral region matches a distinct class of transition.

Why is the visible region so narrow (roughly 400–800 nm) compared to the rest of the spectrum?

The visible region is defined by human retinal photochemistry, not by any special property of electromagnetic radiation itself. Rhodopsin and cone photopigments in the eye happen to undergo a productive cis-trans photoisomerization only for photons in the 400–700 nm range; outside this window, photons either lack sufficient energy to drive that specific isomerization (infrared) or are absorbed before reaching the retina by the cornea, lens, or atmospheric ozone (deep UV). Spectroscopically there is nothing uniquely privileged about 400–800 nm — many analytically important electronic transitions occur just outside it, in the near-UV or near-IR, and are measured routinely by instruments (unlike the human eye) that are not restricted to visible detection.

Why do IR spectrometers report wavenumbers (cm⁻¹) instead of wavelength or frequency directly?

Wavenumber $\tilde\nu = 1/\lambda$ is directly proportional to energy ( $E = hc\tilde\nu$ ), so equal spacings on a wavenumber axis correspond to equal spacings in energy — convenient for comparing vibrational transitions of different bonds on one linear scale. Frequency in Hz would require unwieldy numbers (~10^13–10^14 Hz for mid-IR), while wavelength in micrometers is not linearly proportional to energy, distorting visual comparison of band spacings. Wavenumbers also happen to give convenient two-to-four digit numbers (400–4000 cm⁻¹) for the mid-IR fingerprint region, which is a strong practical reason the convention persists industry-wide.

Does "ultraviolet" always mean the same energy range in every context?

Not exactly — UV is subdivided further in different disciplines. Photobiologists commonly split it into UV-A (315–400 nm), UV-B (280–315 nm), and UV-C (100–280 nm) based on skin and DNA photochemistry, while analytical chemists more often speak of "near UV" (200–400 nm, accessible with ordinary quartz optics in air) versus "vacuum UV" (10–200 nm, requiring evacuated or purged optical paths because atmospheric oxygen and nitrogen strongly absorb below ~200 nm). All of these are conventional subdivisions of the same continuous photon-energy range; the physics (valence-electron transitions) is identical across them, only the exact cutoff wavelengths differ by convention and application.

Why can't a single low-energy photon combine with another to drive a high-energy transition in ordinary spectrochemical analysis?

In principle, two-photon or multiphoton absorption is a real quantum-mechanical process, but it requires photon flux densities (photons per unit area per unit time) far higher than a conventional lamp or even most continuous lasers provide, because the two photons must arrive at the same absorbing center within an extremely short time window to be absorbed "simultaneously." This is why multiphoton spectroscopy is a specialized, high-intensity pulsed-laser technique (used, for example, in multiphoton microscopy) rather than the default mode of absorption spectrophotometers. For essentially all routine UV-Vis, IR, and X-ray spectrochemical analysis, the single-photon resonance condition $E_\text{photon} = \Delta E_\text{gap}$ is the operative rule.

How does the choice of spectral region relate to what "matrix" or solvent a sample can be measured in?

Different regions demand different sample containment because common solvents and window materials absorb strongly in some regions and not others. Water and glass are essentially opaque in much of the mid-infrared, forcing IR spectroscopists to use KBr pellets, ATR crystals, or non-aqueous solvents; by contrast, water is an excellent solvent for UV-Vis work because it is transparent from about 190 nm upward. X-ray measurements largely ignore ordinary solvent absorption because core-electron binding energies vastly exceed anything a solvent's valence or vibrational structure could interfere with, but they instead demand attention to sample thickness and self-absorption effects. Choosing an analytical region therefore constrains, and is constrained by, the physical sample-handling method available.

Is there a sharp physical boundary between "vibrational" infrared absorption and "electronic" visible absorption?

No — the boundary near 800 nm to a few micrometers is a practical convention rather than a sharp physical wall. Some large, low-gap conjugated molecules (certain dyes, semiconducting nanoparticles, or transition-metal complexes with small ligand-field splittings) show genuine electronic transitions extending into the near-infrared (700 nm–2 μm), overlapping the region often loosely called "near-IR" in vibrational overtone spectroscopy. The physically correct statement is always that a photon of a given energy drives whichever transition (vibrational or electronic) happens to have a matching energy gap in that particular molecule — the region label is a convenient generalization, not a universal law.

Why does XRF use characteristic X-ray emission lines rather than absorption, while UV-Vis analysis typically uses absorption?

Both processes are available in principle in every region, but practical sensitivity favors different modes. In XRF, an incident X-ray ejects a core electron, and a higher-shell electron falls to fill the vacancy, emitting a photon whose energy is a sharp, element-specific fingerprint (e.g., Fe Kα at 6.40 keV) against a low background — emission is easy to detect with high signal-to-noise. In UV-Vis, transitions are far more numerous, broadened by vibrational and solvent effects, and easily quantified through Beer's law absorbance measurements against a well-defined incident beam; emission (fluorescence) is also used but is generally a secondary, more specialized measurement in that region. The choice reflects which measurement mode gives the cleanest, most quantitative signal in each region's characteristic linewidths and background conditions.