Data from FTIR spectrometer generally come in a form of ASCII (x,y) pairs (.csv, .tbl, .xy, .asc, .prn, .txt, .dpt, .dat, etc.). There are plenty of spreadsheet software which could be used for production of fine line graphs out of these data, but very few of them can actually export images in the requested resolution for scientific journal publication (300 dpi). Hereby, I would like to propose some code for plotting that kind of graphs. The presented example is part of Cujic et al. (2016) experiment on micro-encapsulation of Aronia melanocarpa EtOH extract. We have used data exported in .prn format, but 'read.table' function should be able to read all potential FTIR data formats. You can download source .prn data from the link below.
library(ggplot2)
library(Hmisc)
directory <- "./data/ftir/" # Define directory where ftir data files are stored
files <- list.files(directory) # Get list of the all files from the directory
lng <- length(files) # Get number of the files
# Set up needed variables to NULL
df <- NULL
tbl <- NULL
names <- NULL
# Create wavenumber vector (x-axis) from the first column of the first file
wave <- read.table(file.path(directory, files[1]), header = FALSE)[, 1]
# Loop for second column in each data file and store in 'tbl' data frame
for(i in 1:lng){
df <- read.table(file.path(directory, files[i]), header = FALSE)[,2]
names <- c(names, sub("^([^.]*).*", "\\1", files[i]))
tbl <- cbind(tbl, df)
}
sq <- seq(from = -35, to = 50*lng, length.out = lng) # Create vector of separator values
tbl <- sapply(1:ncol(tbl),function(x) tbl[,x] + sq[x]) # Separate each ftir for about 50 units along y-axis
tbl <- cbind(wave, tbl) # Bind first column of wavenumbers (x-axis)
colnames(tbl)[-c(1)] <- names # Give column names as original files (may be useful)
tbl <- as.data.frame(tbl) # Define 'tbl' as data frame
# Define image type, dimensions and quality
tiff(filename = "./data/ftir.tiff",
width = 3000, height = 4000,
units = "px", pointsize = 12,
compression = c("lzw"),
bg = "white", res = 300,
type = c("cairo"))
par(mfrow=c(1,1), mar=c(3, 3, 1, 2), las=0)
# Create blank plot
plot(tbl[,1], tbl[, lng], type = "n",
xlab = "", ylab = "", xaxs="i",yaxs="i",
ylim = c(0, sq[lng]), xlim = c(4000, 0),
bty="n", xaxt="n", yaxt = "n")
# Draw ftir line graphs
for(i in 2:(lng-1)){
lines(tbl[,1], tbl[,i], lty = 1)
}
# Axis definition
minor.tick(nx = 10, ny=0, tick.ratio=0.2)
axis(1, cex.axis = 0.8, cex.lab = 0.8, tck = -0.01, mgp = c(3, 0.1, 0))
axis(4, cex.axis = 0.8, cex.lab = 0.8, tck = FALSE, pos = 400, labels = FALSE)
mtext(side=1, text=expression("Wavenumber [cm"^-1*"]"), line=1.5, cex=0.8)
mtext(side=4, text="Intensity [a.u.]", line=-3, cex=0.8)
# Put letters above each ftir chromatogram
lett <- rev(letters[seq( from = 1, to = lng - 1)])
for(i in 2:lng){
text(x = 3851.5, y = tbl[tbl$wave == 3851.5, i] + 10,labels=paste("(",lett[i],")", sep = ""), cex = 1.2)
}
dev.off()
Output has been presented in Figure 1. Please observe carefully two options that were chosen. First, each chromatogram have been separated by 50 intensity units and shifted down the y-axis all of them by 35 units. These scalings are optional and depend on the nature of your chromatograms. Second, wavenumber of 3851.5 have been chosen as optimal for locating labels. Again, it depends on the nature of your samples.
 |
| Figure 1. FTIR spectra of (a) freeze dry chokeberry extract; (b) blank F2 beads; (c) blank F4 beads; (d) F2 beads encapsulating chokeberry extract (e) F4 beads encapsulating chokeberry extract. (Cujic et al., 2016) |
Reference: N. Ćujić, K. Trifković, B. Bugarski, S. Ibrić, D. Pljevljakušić, K. Šavikin (2016) Chokeberry (Aronia melanocarpa L.) extract loaded in alginate and alginate/inulin system, Industrial Crops and Products 86, pp. 120-131.
Source files: https://github.com/Shansh/chromatography/tree/master/ftir
