Intra-Night Optical Variability in AGNs

• INOV is the study of rapid, intra-night optical flux variations in AGNs that reveals minute-scale changes in both jet and disk emissions.
• Differential photometry, combined with robust statistical tests like the F-test, enables detection of variability at the 1–2% level.
• INOV serves as a diagnostic tool to distinguish AGN classes, highlighting blazar jet dominance and more subtle variations in radio-quiet quasars.

Intra-night optical variability (INOV), also called optical microvariability, denotes optical flux changes on timescales of minutes to hours within a single night. In active galactic nuclei, it is used as a compact-region diagnostic because rapid optical changes can probe emission processes in the innermost disk-jet system that remain unresolved by direct imaging. The modern literature treats INOV both as a phenomenological signature—especially of Doppler-boosted jet activity—and as a comparative tool for distinguishing radio-quiet quasars, radio-loud quasars, BL Lac objects, TeV blazars, narrow-line Seyfert 1 galaxies (NLSy1s), and other special AGN subclasses (Gopal-Krishna et al., 2017).

1. Definition, scope, and historical development

The observational concept is straightforward: a target AGN is monitored continuously for several hours, and its optical brightness is compared against nearby field stars recorded on the same CCD frames. This same-frame differential photometry largely removes transparency and airmass fluctuations, making it possible to test for variability at the percent level. In the historical overview from ARIES, the earlier label “optical microvariability” is explicitly said to have given way to “intra-night optical variability” (INOV), and the major matched-sample campaigns typically monitored each AGN for 4–8 hours per night, often around 5–6.5 hours, with detection thresholds of $1\!-\!2\%$ in the best programs (Gopal-Krishna et al., 2017).

By the 1990s and 2000s, INOV had shifted from a phenomenon mostly associated with blazars to a broader comparative diagnostic across the main AGN classes. The ARIES/UPSO programs are described as having been among the first systematic efforts to compare radio-quiet quasars, lobe-dominated quasars, core-dominated quasars, polarization-selected quasars, and TeV blazars under homogeneous observing and reduction procedures (Gopal-Krishna et al., 2017). A major outcome of that tradition was the recognition that strong INOV is not simply a synonym for radio loudness: mild, low-duty-cycle INOV occurs in radio-quiet quasars as well, whereas the most frequent and largest-amplitude INOV belongs to blazars, especially strongly polarized ones (Gopal-Krishna et al., 2017).

This historical trajectory matters because later work on NLSy1s, low-mass AGN, survey-selected quasars, and radio-state transition objects explicitly extends that comparative framework rather than replacing it. INOV is therefore best understood not as a single-source curiosity but as a standardized time-domain observable with class-dependent incidence, amplitude, and physical interpretation.

2. Observational methodology and statistical formalism

Modern INOV work is built on differential light curves (DLCs) between the target and two or more comparison stars, plus one or more star-star DLCs used as control curves. Aperture photometry remains standard, with aperture radius often set to $2\times$FWHM or empirically optimized by minimizing the scatter of the steadiest star-star DLC. Because low-redshift AGN can suffer spurious variability from host-galaxy contamination under variable seeing, many recent studies explicitly inspect the seeing evolution and reject or qualify sessions in which the DLC morphology tracks PSF changes (Ojha, 2022).

A recurring technical issue is that formal IRAF/DAOPHOT errors are underestimated. Several studies therefore adopt an empirical error-scaling factor close to $\eta \simeq 1.5$, while some recent work uses $\eta = 1.54 \pm 0.05$ (Goyal et al., 2013). The standard error-corrected $F$-test is often written as

$$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$

where the numerator is the observed variance of the AGN-star DLC and the denominator is the scaled mean square formal error (Ojha et al., 2020). A more sensitive alternative is the power-enhanced $F$-test,

$$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$

which pools multiple comparison-star DLCs after scaling them to the noise level of the target DLC (Ojha et al., 2020).

For variability amplitude, a widely used estimator is the Heidt–Wagner expression, written in some papers as $Amp$ and in others as $\psi$: $2\times$0 with $2\times$1 and $2\times$2 the extrema of the target light curve and $2\times$3 the representative measurement error (Gaur et al., 2015). Duty cycle is commonly computed with the Romero et al. duration-weighted, redshift-corrected definition,

$2\times$4

where $2\times$5 for a variable session and $2\times$6 otherwise (Gaur et al., 2015).

The statistical choice is not trivial. A benchmark reanalysis of 262 monitoring sessions of 77 AGN showed that the $2\times$7-test produced far too many false positives on star-star DLCs—43 and 59 at $2\times$8 and $2\times$9, against expectations of roughly $\eta \simeq 1.5$0 and $\eta \simeq 1.5$1—whereas the $\eta \simeq 1.5$2-test and modified $\eta \simeq 1.5$3-test yielded 6 and 18, close to expectation (Goyal et al., 2013). That result established the $\eta \simeq 1.5$4-test family as the preferred standard for comparative INOV work.

3. Class dependence across the major AGN populations

A homogeneous benchmark for class-dependent INOV came from the reanalysis of 262 intra-night sessions of 77 AGN spanning six prominent classes (Goyal et al., 2013). Using the validated $\eta \simeq 1.5$5-test, the reported duty cycles were:

AGN class	INOV DC	DC for $\eta \simeq 1.5$6
RQQs	10%	6%
RIQs	18%	11%
LDQs	5%	3%
LPCDQs	17%	10%
HPCDQs	43%	38%
TeV blazars	45%	32%

This hierarchy is one of the central empirical results of the field. Radio-quiet quasars and lobe-dominated quasars show mild, infrequent INOV; radio-intermediate quasars and low-polarization core-dominated quasars are only modestly more active; high-polarization core-dominated quasars and TeV blazars occupy the high-duty-cycle, high-amplitude end (Goyal et al., 2013).

A focused study of core-dominated radio quasars sharpened the point that radio core dominance and relativistic beaming are not, by themselves, sufficient for strong INOV. In a combined sample of 21 CDQs monitored on 73 nights, the duty cycle was $\eta \simeq 1.5$7 for low-polarization CDQs and $\eta \simeq 1.5$8 for high-polarization CDQs; restricting to strong events with $\eta \simeq 1.5$9, the corresponding values were $\eta = 1.54 \pm 0.05$0 and $\eta = 1.54 \pm 0.05$1 (Goyal et al., 2012). The paper’s conclusion was explicit: relativistic beaming is normally not a sufficient condition for strong INOV and a high optical polarization is the other necessary condition (Goyal et al., 2012).

The historical synthesis from ARIES is consistent with that picture. It reports mild INOV in radio-quiet QSOs, with amplitudes up to $\eta = 1.54 \pm 0.05$2 and duty cycle $\eta = 1.54 \pm 0.05$3, and emphasizes that similarly mild INOV is exhibited even by the vast majority of radio-loud quasars, including many with powerful relativistic jets. The solitary outliers are blazars, especially the strongly polarized subset, which frequently exhibit pronounced INOV (Gopal-Krishna et al., 2017). The class dependence of INOV is therefore best described as a hierarchy of jet dominance, beaming, and optical synchrotron prominence, not a binary radio-loud/radio-quiet split.

4. NLSy1 galaxies as a critical INOV laboratory

NLSy1 galaxies have become one of the most important arenas for INOV studies because they combine Seyfert-like optical spectra with, in some cases, unmistakably blazar-like jet signatures. A first systematic comparison of NLSy1 subclasses found duty cycles of $\eta = 1.54 \pm 0.05$4, $\eta = 1.54 \pm 0.05$5, and $\eta = 1.54 \pm 0.05$6 for $\eta = 1.54 \pm 0.05$7-ray-loud, $\eta = 1.54 \pm 0.05$8-ray-quiet radio-loud, and radio-quiet NLSy1s, respectively; the corresponding mean amplitudes were $\eta = 1.54 \pm 0.05$9, $F$0, and $F$1 in the tabulated sense (K. et al., 2016). The same study further reported that high-polarization radio-loud NLSy1s show somewhat higher duty cycles and amplitudes than low-polarization ones, reinforcing the polarization–INOV connection already known from quasar studies (K. et al., 2016).

A complementary comparison of X-ray- and $F$2-ray-detected NLSy1s reached a closely related conclusion. The full X-ray-only sample had an INOV duty cycle of $F$3, whereas the $F$4-ray sample reached $F$5; however, once the X-ray sample was subdivided by radio loudness, the 13 radio-quiet X-ray NLSy1s had $F$6 duty cycle while the 5 radio-loud X-ray NLSy1s had $F$7, very similar to the $F$8-ray sample (Vineet et al., 2020). The paper therefore concluded that radio loudness level is the prime factor behind INOV detection and the pattern of the high-energy radiation plays only a minor role (Vineet et al., 2020).

The same logic appears in VLBA-based jet classifications. In a systematic comparison of 15 jetted and 8 non-jetted radio-loud NLSy1s, the conservative $F$9-test yielded an INOV duty cycle of $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$0 for the jetted subsample and $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$1 for the non-jetted subsample; among the jetted objects, the $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$2-ray-detected members reached $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$3, while the non-$$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$4-ray jetted members again showed no INOV (Ojha et al., 2022). This result was interpreted as evidence that mere jet presence is not enough: relativistic beaming plays a significant role, especially in high-accretion systems whose optical non-thermal emission can be diluted by the steadier accretion disc (Ojha et al., 2022).

At class level, $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$5-ray NLSy1s behave much more like blazars than like ordinary Seyferts. A uniform 15-source, 36-session campaign found a duty cycle of around $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$6 conservatively and around $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$7 with the enhanced $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$8-test; the mean INOV amplitude over variable sessions was $$F_{1}^{\eta} = \frac{\sigma^{2}_{(q-s1)}}{\eta^2 \langle \sigma_{q-s1}^2 \rangle}, \qquad F_{2}^{\eta} = \frac{\sigma^{2}_{(q-s2)}}{\eta^2 \langle \sigma_{q-s2}^2 \rangle},$$9, and the authors argued that these values may require upward revision once dilution by host-galaxy and accretion-disc light is corrected (Ojha et al., 2020). A plausible implication is that the observed optical INOV of $F$0-NLSy1s is a lower limit on the intrinsic jet variability.

The NLSy1 regime has also broadened in unexpected directions. A sample of seven apparently radio-quiet NLSy1s that had shown recurring 37 GHz flaring but no clear JVLA jet signatures displayed an average duty cycle of $F$1, similar to that reported for $F$2-ray NLSy1s (Ojha et al., 2024). That paper inferred relativistic jets in the sample but also noted that magnetic reconnection in the black-hole magnetosphere remains a viable mechanism for the observed INOV (Ojha et al., 2024). In that sense, NLSy1s have turned INOV into a direct test of hidden, intermittent, weak, or otherwise ambiguous jet activity.

5. Phenomenology: amplitudes, timescales, chromaticity, and extreme events

The phenomenology of INOV is richer than duty cycle alone. Intensive quasi-simultaneous $F$3 monitoring of BL Lacertae over 38 nights produced 113 light curves, with significant INOV on 19 nights and a maximum intranight amplitude of $F$4 in the $F$5 band on 6 July 2010, when the source changed by about 0.3 mag in less than 3.5 hours (Gaur et al., 2015). The same study found no evidence for periodicities or characteristic variability timescales, strong interband correlations with lags constrained to be shorter than the $F$6 min sampling interval, a general tendency for larger amplitudes at higher optical frequencies, and significant bluer-when-brighter relations in 5 of 13 suitable intranight observations (Gaur et al., 2015).

Systematic TeV-blazar monitoring established an important counterpoint to expectations from extreme jet Lorentz factors. In a 22-object sample monitored on 116 nights for a cumulative 677 h, with typical cadence around 10 min and sensitivity to $F$7–2% fluctuations, no clear variability feature on timescales substantially shorter than 1 h was found (Gopal-Krishna et al., 2011). The overall $F$8-test duty cycle was $F$9, dropping to $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$0 for $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$1, and the well-known LBL/HBL contrast persisted even within the TeV sample, with duty cycles of $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$2 and $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$3, respectively (Gopal-Krishna et al., 2011). This suggests that minute-scale optical events are not the norm even in AGN whose TeV phenomenology implies ultra-relativistic inner jets.

Some NLSy1 case studies occupy the opposite end of the timescale distribution. PMN J0948+0022 was the first radio-loud NLSy1 in which violent optical INOV of roughly $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$4 mag within several hours was reported in both $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$5 and $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$6 bands (Liu et al., 2010). In a three-source campaign of $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$7-ray loud NLSy1s, duty cycles of $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$8 with the $$F_{\rm enh} = \frac{s_{\rm qso}^2}{s_{\rm stc}^2}, \qquad s_{\rm stc}^2=\frac{1}{(\sum _{j=1}^k N_j) - k}\sum _{j=1}^{k}\sum _{i=1}^{N_j}s_{j,i}^2,$$9-statistic and $Amp$0 with the $Amp$1-statistic were reported, and mini-flares as short as 12 minutes were detected (Paliya et al., 2012). Those results placed at least some $Amp$2-NLSy1s squarely in the blazar regime of optical microvariability.

An especially revealing single-source case is the non-jetted radio-loud NLSy1 SDSS J163401.94$Amp$3480940.2. In one of two 3-hour monitoring sessions, the source showed a sharp optical flare with mean peak-to-peak amplitude $Amp$4; the light curve brightened by $Amp$5 in 6.01 minutes and, under a conservative 50%-disk/50%-jet decomposition, implied a flux-doubling time of $Amp$6 minutes (Ojha, 2022). The flare survived re-extraction at the original 100 s cadence, where it was resolved by four points rather than two, and the paper interpreted it as blazar-like variability in a source previously classified as non-jetted from VLBA imaging (Ojha, 2022). This is a paradigmatic example of INOV revealing relativistic activity where radio morphology alone remained ambiguous.

6. Interpretation, controversies, and emerging directions

The dominant interpretation of strong INOV is jet-based, but the literature repeatedly shows that the mapping is not one-to-one. Strong, frequent INOV is empirically associated with Doppler-boosted synchrotron jets, yet radio loudness alone does not guarantee it, and even parsec-scale jet detection is not always enough. Conversely, a single VLBA image lacking a resolved jet does not imply the absence of jet-dominated variability. In the case of SDSS J163401.94$Amp$7480940.2, the optical flare was reconciled with the VLBA non-detection through possible compactness, temporal intermittency, low sensitivity, a weak or frustrated jet, synchrotron self-absorption, or free-free absorption in dense circumnuclear gas (Ojha, 2022). This suggests that INOV can trace unresolved or hidden jets that radio morphology misses.

The survey era has further complicated any simple blazar-only view of strong INOV. A ZTF-based study of 53 blazars and 132 radio-quiet quasars, matched in the redshift–magnitude plane and analyzed over 156 and 418 intranight sessions, confirmed INOV in 6 blazar sessions but also in 3 RQQ sessions; strikingly, all three confirmed RQQ events had $Amp$8, with amplitudes of 13%, 14%, and 15% (Negi et al., 2023). The paper concluded that a blazar-like INOV level can also be attained by RQQs, albeit very rarely (Negi et al., 2023). A plausible implication is that rare micro-jet activity, occultation-like dips, or other compact non-thermal processes may occasionally operate even in genuinely radio-quiet nuclei.

INOV has also become relevant to transient and evolutionary problems. In the quasar J0950+5128, identified as a candidate for transition from radio-quiet to radio-loud state, six dedicated $Amp$9-band sessions each longer than 4 h yielded no sustained INOV down to the 1–2% level, although a $\psi$0-mag spike lasting $\psi$1 minutes was seen on one night; archival ZTF data, however, showed significant INOV in two of three January 2019 sessions, around the epoch when the radio-loud transition was recognized (Chand et al., 2024). This suggests that INOV associated with radio-state transitions may be episodic rather than persistent.

The mass scale of INOV-active nuclei has also expanded downward. A representative sample of 12 low-mass AGN with $\psi$2, monitored in 36 sessions, showed significant INOV in 8 sessions and another 2 probable-variable sessions, implying duty cycles of $\psi$3 and $\psi$4 if probable variables are included; all definite-variable sessions had $\psi$5 (Gopal-Krishna et al., 2022). That study concluded that the INOV level is statistically comparable to that observed for blazars and $\psi$6-ray detected NLSy1s (Gopal-Krishna et al., 2022). This suggests that blazar-level optical microvariability can be sustained by central engines with black holes near the upper end of the intermediate-mass range.

Taken together, these results establish INOV as both a diagnostic and a cautionary observable. It is one of the most sensitive indirect probes of relativistic beaming in AGN, yet its observed amplitude and duty cycle can be strongly modulated by optical polarization state, host-galaxy contamination, accretion-disc dilution, cadence, duration of monitoring, and source classification based on incomplete radio information. The most secure generalization is therefore not that INOV uniquely identifies blazars, but that its strongest, fastest, and highest-duty-cycle manifestations delineate the regimes in which compact non-thermal jet activity dominates the optical continuum.