SOL 101 › Lab

Lab 101-A — Sunlight versus Lamps with a Diffraction-Grating Spectroscope

Goes with: Units 3, 4, 5, 7. Time: 2–3 hours (build 45 min, measure 1 h, analyse 1 h). Cost: about $25–$45 without the optional spectrometer (estimate).

Goal

Build a simple spectroscope, photograph the spectra of indirect sunlight, an incandescent/halogen bulb, a warm-white LED and a fluorescent/CFL lamp, turn the photos into rough intensity-vs-wavelength curves, and compare them with the AM1.5G reference spectrum and with Planck curves.

What you will learn

How a grating separates wavelengths ($d\sin\theta = m\lambda$).
Why sunlight is broad and smooth (blackbody-like with absorption dips) while LEDs and fluorescent lamps are spiky.
Why lamps are poor solar simulators.
That cameras see some near-infrared that eyes do not.

Shopping list

Item	Spec / example	Qty	Approx. cost (estimate)	Where
Diffraction grating	linear, 500 or 1000 lines/mm, film sheet or slide (e.g., Rainbow Symphony or Edmund Optics holographic grating film)	1 sheet/slide	$5–15	Amazon, Edmund Optics, science-kit suppliers
Box	shoebox or cereal box, plus black paper/paint	1	$0–5	household
Slit material	two single-edge razor blades or aluminium foil + craft knife	1 set	$3–5	hardware store
Black tape, glue, ruler	—	—	$5	household
Lamps	E26 incandescent or halogen bulb (if obtainable), warm-white LED bulb, CFL or fluorescent tube	1 each	$10–20	hardware store
Lamp holder	clamp lamp with E26 socket	1	$10	hardware store
Phone with camera	manual exposure app helps (fix exposure/white balance)	1	—	—
Optional calibrated spectrometer	low-cost USB/hobby spectrometer (availability varies; e.g., open-hardware Public Lab designs, or a commercial unit)	1	$50–300 (not price-checked)	Public Lab / vendors

Safety

Never look at the Sun

Never point the spectroscope or camera at the Sun and never look at the Sun through the grating. Use sunlight reflected from white paper or a white wall, or blue sky away from the Sun.

Halogen and incandescent bulbs get very hot (hundreds of °C at the glass for halogen): use a clamp lamp, never touch the bulb, keep away from paper.
Razor blades: tape the back edges; dispose in a sharps container.
CFLs contain small amounts of mercury — don't break them; if one breaks, ventilate and follow EPA clean-up guidance.

Build the spectroscope

Black out the inside of the box (paint or black paper) to cut stray light.
At one end, cut a 2 cm × 3 cm window and tape two razor blades edge-to-edge across it, leaving a slit about 0.2–0.5 mm wide (a business card is ~0.3 mm thick — use it as a spacer).
At the opposite end (≥ 25 cm away), cut a window for the grating. Mount the grating so its lines run parallel to the slit. Tilt it so the first-order spectrum appears at roughly 20–30° off the straight-through direction (for 500 lines/mm: 400–700 nm → 11.5°–20.5°; for 1000 lines/mm: 23.6°–44.4°).
Hold the phone against the grating window, find the spectrum, and tape the phone mount in place so every photo has the same geometry.

Calibrate the wavelength scale

A fluorescent or CFL lamp has sharp mercury lines at 404.7, 435.8, 546.1 and 577/579 nm (standard values), plus europium/terbium phosphor peaks near 611 nm and 545 nm. Photograph the CFL spectrum first: 1. Open the photo in any image tool (or Python) and note the pixel column of the 435.8 nm (blue) and 546.1 nm (green) lines. 2. Fit a straight line wavelength = a × pixel + b. Check it predicts ~577–579 nm for the yellow doublet within ~5 nm.

Procedure

Fix camera exposure, ISO and white balance (manual mode) so photos are comparable. Avoid saturating (clipped) pixels.
Photograph, in a darkened room: CFL (calibration), incandescent/halogen, warm-white LED, then outdoors: sunlit white paper (not the Sun!) and blue sky.
For each photo, take a dark frame (lamp off) for background subtraction.
Record lamp type, wattage, colour temperature on packaging, and time/weather for sunlight.

Analysis

Python sketch (use Pillow + numpy):

from PIL import Image; import numpy as np
img = np.asarray(Image.open("halogen.jpg").convert("RGB"), float)
dark = np.asarray(Image.open("dark.jpg").convert("RGB"), float)
band = (img - dark)[y0:y1, :, :].sum(axis=(0, 2))   # rows covering the spectrum
wl = a * np.arange(band.size) + b                     # your calibration

1. Plot intensity vs wavelength for each source on one graph (normalise each to its own maximum). 2. Overlay the AM1.5G curve (download the G173 table — Unit 8) normalised the same way, and Planck curves for 2700, 3200 and 5772 K. 3. Important caveat: a phone camera's sensitivity varies strongly with wavelength (and an IR-cut filter removes most light beyond ~680 nm), so your curves are camera-weighted, not true spectra. Compare sources with each other; don't over-interpret absolute shapes. The optional calibrated spectrometer removes this limitation.

What to measure / record

Peak wavelength of each source (after calibration).
Width of the spectrum; presence of lines (CFL), two humps (white LED: blue pump ~450 nm + yellow phosphor), smooth rising curve (incandescent).
Relative blue-to-red ratio (e.g., 450 nm / 650 nm intensity) for each source.
Dark bands in the sunlight spectrum: with a good slit you may see the Fraunhofer lines (e.g., sodium D near 589 nm, hydrogen near 656 nm) — evidence of absorption in the Sun's atmosphere.

Expected results

Source	Expected look
Sunlit paper / sky	continuous, fairly flat across visible; sky noticeably bluer (Rayleigh)
Incandescent/halogen	continuous, rising steeply toward red (peak is in the IR: Wien ≈ 0.9–1.1 µm)
Warm-white LED	narrow blue peak ~450 nm + broad yellow-orange hump; little deep red, no IR
CFL	discrete lines (Hg) and narrow phosphor peaks

How to interpret

The incandescent curve is a Planck curve's short-wavelength tail — much more red than blue. Most of its output is invisible IR: a poor solar simulator.
LED and CFL spectra have gaps; a solar cell's current under them would depend strongly on its spectral response — which is why professional simulators are graded for spectral match (IEC 60904-9).
Sunlight's spectrum is the reference that everything is compared to; its smoothness reflects a ~5800 K blackbody source, and any dark lines are atomic absorption.

Extensions

Put a silicon cell (Lab 101-B) under each lamp at equal lux reading and compare short-circuit current — the cell responds to photons (including IR) that a lux meter ignores.
Use an IR remote control: the camera may show its ~940 nm LED as purple-white even though your eye sees nothing.

Sources

Mercury emission lines: NIST Atomic Spectra Database (standard values quoted).
Grating equation: Hecht, Optics.
ASTM G173-03 (comparison spectrum).

← previous next →